May 11, 2023 • 45 mins
Supply chain design has traditionally been conducted with only one dimension in mind: cost minimization. The conventions of supply chain design date largely from the 1990s. But major technological strides have been made in the years since, allowing for much more complex considerations to be taken into account. So most companies, by using this obsolete approach, are putting themselves at a competitive disadvantage.
Today's episode features Milena Janjevic, Research Scientist at the MIT Center for Transportation & Logistics, and Mike Bucci, a Senior Director for Solutions Delivery at Coupa Software. Milena and Mike discuss how they're seeing the field evolve, touching on four big opportunities to take supply chain design from a cost-minimization exercise to an engine to drive value creation:
- Extending the scope of supply chain design
- Incorporating tactical decision making
- Accounting and planning for risk and resilience
- Adapting new technologies and business models
For more, learn about the Supply Chain Design Initiative at the MIT Center for Transportation & Logistics.
Host: Arthur Grau Executive Producer: Benjy Kantor Marketing Writer & Producer: Dan McCool Sound Editor: David Benjamin Sound
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
(upbeat music)
- Welcome to MIT Supply Chain Frontiers
from the MIT Center forTransportation and Logistics.
Each episode featurescenter researchers and staff
or experts from the fieldfor in-depth conversations
about business education and beyond.
(upbeat music)
(00:22):
- Thank you for joining today's,
"MIT CTL Supply Chain Frontiers," podcast.
So happy to have MikeBucci and Milena Janjevic.
Milena is a researchscientist here at MIT CTL.
And Mike, can you tell usa little bit about yourself
and your role?
- Yeah, sure thing.
Mike Bucci, I work forCoupa, formerly LLamasoft
(00:43):
as a few years back.
Been largely in the services,professional services group
during that time.
Mostly working with companies
around the world in all industries.
Kind of helped them understand how
to leverage the Coupa Software
to solve their supplychain design problems.
So I've been fortunate towork with many companies
(01:04):
around the globe and looking forward
to sharing some of thatexperiences here today.
- Excellent.
And Milena, can you tell usa little bit about your work
and how it relates to supply chain design
and the work that Mike is doing?
- Yes, thanks for having me here.
So I am leading the supplychain design initiative,
which is basically looking at the ways
in which we can improvethe decision making process
(01:28):
around supply chain design.
So here at MIT CTL, we do a variety
of industry sponsored researchand educational programs
with the aim of helpingcompanies learn how
to better address theirsupply chain design problems.
- Great, great, great.
So for today we're gonna betalking a little bit about sort
(01:49):
of old world or past use supplychain design and what some
of your research and somewhat new practices are
that you're doing in asupply chain design space.
So you recently co-publisheda white paper called,
The New Competitive Edge Analytics
Driven Supply Chain Design.
And in the white paper you mentioned
(02:10):
that most companies are still
using standard supplychain design practices
that have been in placefor around 30 years.
What is the traditionalsupply chain design approach?
- From the Coupa's side, Imean really what we've seen
over the years and therehas been a transition,
but yeah, the traditionalapproach and practice here
for supply chain design has really been
(02:31):
around more event-based orepisodic kind of processes
that occur at some schedule,
but usually less frequently every year
through two to three years.
And the idea behind that is really
to reassess a significancesupply chain problem,
strategic supply chain problem.
It could be a DC network design problem.
(02:52):
It could be productioncapacity planning problem,
but it's more of thatevent-based project or process.
And there's really a verysignificant heavy lift
to go find the data, collectthe data, cleanse the data.
You know, get organizationalbuy-in and the process.
Provide recommendations andthen go implement the solution.
(03:14):
And so that's kind of aprocess that reoccurs,
but it's kind of again, thismore of event-based problem.
And the focus there istypically on minimizing cost.
It includes, you know,obviously balancing service
and inventory, but it'sjust that general problem.
But again, it's more ofthat event-based activity.
- Does the periodicity or the time in
(03:36):
between updates affect thedifficulty of you to get the data
in a situation like that orto get accurate information?
- So I'd say that when you are redesigning
your supply chains every fiveyears or 10 years, the level
of granularity that youwill be able to incorporate
in those supply chain studies,it's typically very low.
(03:58):
So typically companies will aggregate data
at a very high level and that will lead
to all kinds of approximationsthat will ultimately result
in a big gap betweenthe expected performance
of your supply chain
and then the realized performance
once you implement a new design.
(04:19):
And some of that has been, Iwould say has resulted because
of simply the computational power
that was limited in the past.
But now we see that the data availability
has significantly increased,
but the supply chain design practices
have not necessarily caught up with that.
(04:40):
- Right. Anything elseto add on that point?
Like as far as the length of time between?
- Yeah, yeah, I think as Milena said,
I mean historically we'vebeen somewhat constrained
by computational power
and other things arecertainly looking forward
to what we see in the future.
What we see coming intoplay now is kind of
(05:00):
that more frequent analysisstudy which brings a whole bunch
of pros and cons related to that.
But we can talk through that.
- Yeah, well I mean let's talk about that.
So before we get to the pros,let's imagine that it's 1995
and every five years we'redoing a supply chain design.
The cons to that that I heard are
(05:20):
that you're just, you'regonna have highest level data,
you're not gonna have any granularity.
Are there any other consto that length in between?
- So one big industry trend
that we see is basically amuch higher pace of change.
And that's gonna beboth on the demand side
and on the supply side.
(05:43):
Companies are facingincreasing customer demands
and the need to readapt the supply chains
in much more frequentmatter to basically cater
to those increasing demands.
And that requires tohave a much higher speed
of adaptation of their
supply chain designs.
(06:04):
- Yeah, I mean thinking back,
as Milena said, I mean if we go
back decades here, thecomputational complexity
or the challenges we had forcomputational power, horsepower
as well as data collection, right?
Because systems were dispersed,uncentralized, unstructured
in many ways to try toconsolidate the data
and effectively leverage itfor these kind of studies,
(06:26):
that was largely the challenges.
So today we obviously havea significant difference,
right, where as Milena said,
supply chains are much morediverse, both from a supply
and demand perspective,there's more continual change
and the frequency of adaptinga supply chain is required
to be much more frequent than in the past.
(06:48):
And all that requires, of course
than the infrastructure around that.
So I think the newer landscape is
that we not only have acomputational capability, we
also have the systeminfrastructures to be able
to pull the data morefrequently and more efficiently
to kind of do thesestudies more frequently
(07:08):
and therefore leverage thoseresults obviously and react
to them much more quickerthan we could in the past.
- Great.
And this is gonna lead
to the conversation aboutsupply chain design.
Just before that, youmentioned consumer demand.
So people want things now.
They want things in theformat, they want 'em
on their front door or theywant 'em in a box somewhere
(07:28):
or maybe they want 'em delivered
to their summer home or whatever it is.
Is there a B2B demand that's changing?
I think is one question that I have.
And then what are some of the other trends
that are pushing towardsthis more tight timeframe
and advanced supply chain design models?
- So I would say thatcustomer centricity has become
a key element and a key sourceof competitive advantage
(07:51):
and both into B2C and B2B realm.
And that's, as Mike mentioned,that's not something
that is typically covered
by the traditional supplychain design methods,
which are really focusedon this cost minimization.
And so whether it's in theB2C or B2B space, we see
that delivery lead time
and delivery reliability arebecoming key order winners
(08:15):
and cannot be ignored fromsupply chain design studies.
- And from the Coupa side,
are you seeing peoplechanging their sourcing models
or is there anythinghappening with globalization
or does that also influence this process?
- Yeah, I would say absolutely.
I mean, we all know that supplychains have been diversified
(08:35):
and global now in nature.
And as we've seen obviously inthe pandemic and other things
along the last severalyears, one disruption along
that supply chain canhave a significant impact
on your entire system.
And so all of thesequestions now that were maybe
in the past more long-term questions
(08:57):
that we could answer once and not have
to revisit them frequentlynow they are questions
that we need to revisit often because
of all the things Milenadescribed, customer behavior
as well as on the supplychain, all the changes
that are occurring on the supply side.
I would also say that the need
to reassess these strategic decisions
more frequently leads kind of a shrinkage
(09:20):
between the strategic andtactical level decisions
that are occurring in companies.
And so now you're makingstrategic decisions
more frequently, maybe quarterly,
and you're making tacticaldecisions maybe monthly
or quarterly.
And now there's kind of a connection there
that we were trying to capture.
And that's one thing we seea lot with the companies
that we're working with is
(09:40):
that though there's a sniff overlap
between those kinds of questions
and those time horizons were typically
were separate kind ofquestions that were asked.
- I think we're understandingthe challenge now.
We know where the market is sitting,
we know where business players are sitting
and what they're up against.
And so in the white paper you mentioned
that there are four opportunities,primary opportunities
(10:01):
that companies and and managers can take
to reimagine the designof their supply chains.
And they are extending the scope,
incorporating the tactical.
Accounting for risk andplanning for resilience
and adopting new technologiesand business models.
So I'd love to get a coupleminutes in on each of these
(10:22):
to find out more about what they mean
and maybe what they mean on the ground
and what you've beenseeing with the people
that you've been working with.
So let's start with extending the scope.
What do you mean by extending the scope?
- Yeah, so the first andmost immediate opportunity
that we see here isextending the objectives
that we are considering inour supply chain design.
(10:44):
So we've already mentioned thattraditional methods focused
on cost minimization on physical structure
of a supply chain.
And that's a very limiting wayof considering supply chains
and supply chain design inthe contemporary market.
And so supply chain designshould have a much broader scope,
(11:05):
which basically focuses onlong-term value creation
and basically consider theinteractions that happen
between the choices we makein the supply chain design
and our ability to generatevalue, increase market share,
and generate revenue.
So for example, one keyquestion in supply chain design
(11:27):
that we have been answering
for the past 30 yearsis where should I look
at my facilities?
How many of thosefacilities should I have?
And if you have purely acost based perspective,
you know the answer's probablygoing to be to centralize
as much as you can andto gain some economies
of scale, have centralizedinventory with lower cost.
(11:49):
But if you consider thisidea of customer centricity
and the fact that the market share
that you can capture willdepend on your ability
to actually serve your clients
in a reliable and fast fashion,then the answer is going
to be completely different.
I do not see that a lotof companies are currently
(12:09):
using approaches that fully account for
that aspect of value creation.
- Yeah, I think Milena covered a lot
of the items I would mention.
I would also say that, you know,
as far as extending thescope, I mean historically
as Milena kind ofoutline, if we're looking
at a supply chain problem,if we're for example looking
at a DC a distribution networkkind of study, we'll put
(12:33):
in the distribution network, some
of the customers maybe alittle bit on the sourcing,
but now we have the abilityto go further back, right?
Maybe we can pick up somemore supplier detail.
We may even include options
of supplier direct kindof shipments, other ways
that we can kind of extend the breadth
of the scope of the supply chain problem.
(12:53):
That's one thing.
And then of course addingin other components or costs
or considerations to the model as well.
Milena obviously discussedthe idea of being closer
to the customer, the impact onhow that can have on demand.
We of course can incorporatethings like taxes and duties
and other kind of new governmentregulations as well as far
as understanding how oursupply chain design should be,
(13:16):
sustainability is anothercomponent as well, right?
Companies are moreconcerned with their CO2
or total emissions and other components.
We can incorporate that aswell kind of into the problem
and evaluate that in conjunctionwith cost, et cetera.
- So that's a little bitabout extending the scope
and broadening what youtraditionally would consider
(13:36):
as a supply chain design problem, right?
So tell us a little bitmore about what you mean
by incorporating the tactical?
- Yeah, I'll start here.
I mean, I think incorporatingthe tactical really is again,
this combined or shrinkingof the difference
between the strategicand tactical decisions.
And in all the work thatwe do with our clients
and customers that we workwith, we see when we try
(13:58):
to address both the strategic question
and the tactical questions,you know, we estimate
that 75% of the datarequirements are the same.
And so why not take advantageand get a multiplier of
that effort to solve bothkinds of problems and have
that same kind of data foundation
where there's not discrepancies
between the tacticaldecision making process
(14:19):
and the strategic process.
So we see a lot of convergence there
of incorporating tactical level detail
into our strategic models.
We're kind of overlapping
the two kind of decision making processes.
And examples of thatwould be kind of things
like an S&OP process,short-term, rough cut,
capacity planning,replenishment inventory planning
(14:42):
kind of policies.
You know, how can we address those more
in our supply chain designkind of problem optimizations.
- One thing that I wouldadd to that is that
from the modeling perspective
and data collectionperspective, the incorporation
of these tactical decisions will
often require a certain effort.
(15:04):
And the idea is not
that every company shouldincorporate each type
of tactical decision with thesame level of granularity,
but really identify those area
that are key for their value creation.
So if I am competing in thelast mile space, I probably need
(15:27):
to have a much more granularand precise integration
of my routing decisionsand my inventory decisions
in the last mile thanif I am a manufacturer
that is mainly going to be looking
at production planning decisions
as a key driver of their advantage.
And so this idea that wecan have a single model
(15:50):
and a single approach
that would fit alldifferent industry context
and companies is now outdated.
- Excellent. I like that, I like that.
So not only is it tactical,
but the tacticalness isgonna depend on where you are
in the supply chain, what your business is
and that sort of thing.
(16:11):
So the next of the fouropportunities that you recognize
in the white paper is to account for risk
and plan for resilience.
And I think this is on everybody's mind
after the last few years.
So can you tell us moreabout accounting for risk
and planning for resilience?
- So yeah, so I think the firststep here is for companies
to basically map out the different sources
(16:35):
of risk, differentsources of vulnerability
and to characterize those.
And there we need to recognize
that when we talk about risk,there are different categories
of risk that will be accounted for
in a very different mannerin our supply chain design.
So if I'm talking about,you know, natural demand
(16:57):
that just comes from somevariations in the markets,
that's very different than a risk linked
to a major natural disaster or pandemic.
And so the way I will account for that
in my supply chain design and
in my tools should bevery, very different.
And so I think that
(17:18):
at this point companiesare very much aware
of the requirement toincorporate the risk,
but they don't necessarilyhave the right tools to employ
and address each type ofrisk that they might face.
- Yeah, to add to what Milena said,
I mean what we're seeingis historically companies
(17:40):
will run scenarios fortheir supply chain design
or tactical problems and they'lldo a few sensitivities of,
well what if TransportationCostco up or down 5%.
But we're seeing obviouslythose sensitivities
increase significantly atwo to three x increase
in number of scenarioscompanies are running today.
Partially due to thisrisk related problem.
(18:03):
So the first thing we see isjust running more scenarios
to look at the resiliency of your network
to different changes.
They can be maybe their cost changes
or they could be changes to constraints.
I have a certain port volume
that I'm expecting to move through.
What if that port volumewere to decrease by 50%?
What would my supply chain,how would my supply chain react
(18:25):
and what is the impactof that on my network?
A second theory to that aswell is we're now able to pull
into some of these models risk scores
and risk metrics, you know,from external sources.
And so by pulling that in someways we can kind of put that
into the objectivefunction or we're trying
to minimize certain amountof risks in our supply chain,
(18:46):
or at least report that out as far
as certain nodes, maybe we'reflowing 95% of our volume
through a certain node.
We wanna decide to highlightthat is a potential risk
because we were kind of single sourced
or single node kind of constrained
if there was a risk eventrelated to that node.
Also, again, based on thoseexternal risk metrics,
(19:10):
we can provide some risk scorefor our supply chain design,
you know, based upon a supplier's risk,
based on differenttransportation risks, et cetera.
And so we're seeing that more and more
in the kind of themodels that we're running
and the requirements our customers
are having regarding risk and resilience.
- I think what's fascinating about this
then is it kind of leads us directly
(19:31):
into the fourth opportunitybecause I'm sensing
that people are maybe collaborating more
or maybe they're gettingdata from resources
that didn't exist beforethat are public or
that are, you know, able to be purchased
and maybe there are new ways
of analyzing and new technology.
So can you tell us aboutthe fourth opportunity
to adopt new technologiesand business models?
(19:53):
- Yeah, I'll start here.
So for what we're seeing,
and you're right on there,Arthur, I mean as far
as what we're seeing is, again,risk is one good example.
We're able to maybeincorporate external data more,
consider that, or atleast, you know, define
that in our solutions.
I would say a couple things here in play,
of course as you mentioned,it's this big data, right?
(20:14):
There's a lot more data we can capture
and bring into our analysis.
The cloud infrastructure iscertainly helping us there
as well, right?
With more information around the cloud,
we can do some interconnectivity.
It's easier to get datathan it was in the past
from different sourcesand different systems
within a company and externally.
AI machine learning playinto that as well, right?
(20:36):
Now we have models wherewe're doing the optimization
or the simulation, butmaybe on the front end
and the back end of thatwe're doing machine learning
and AI to kind of tease outother things in the solution
to better provide results tothe customer, which some of
that can be related to thatrisk and resiliency as well.
And I even think furtherwe're seeing situations where
(20:57):
that AI machine learningcan actually look for things
in our supply chain
that may be I'll saycognitive blind spots maybe
to the modeler, but wedon't really see that issue
or we don't see that problem.
But machine learning orAI can kind of pick up
on potential issues with our supply chain.
It could be risk related orit could be kind of trends
(21:19):
and costs and other thingsthat we're seeing over time
in their networks.
So those are some examples
where I think the newtechnologies are kind of morphing
into this kind of supplychain design problem.
- I would add one thing to that is
that the digital transformationthat Mike has just mentioned
(21:39):
also enables a new types oforganizational relationships
between differentcompanies and relationships
that now include actors thatwere previously not part
of our supply chain ecosystem.
One example of that is the use
of crowdsourced resources,it being transportation
(22:01):
or warehousing and basicallyeverything that is targeted
towards an on-demand useof the resources rather
than investment in assets,
which is a really interestingmitigation strategy
that can build higher resilienceinto your supply chain.
When we talked about thetraditional approaches
(22:22):
to the supply chain designa day would typically assume
a very classical supplierbuyer relationship
that did not reflect these
new complex organizationalrelationships, did not,
for example, incorporate the ways
in which we would sharerevenue, share costs
and share risks amongthose different parties.
(22:44):
So that's another opportunity that we see.
It's extending basically the way
that we represent theorganizational structure
of the supply chains to take advantage
of these new business models
that have arised in the last few years.
- So we've seen the four opportunities are
to extend the scope ofthe supply chain design
(23:06):
to incorporate tactical, moregranular information and data.
Accounting for risk andadopting new technologies.
Do you have any examples of many of these
that you wanna sharefrom your recent work?
- One would be a miningcompany that we've worked with
for many years and they're doing a lot
of strategic analysis,even some tactical studies
(23:28):
on their capacity planning.
But recently they wantedto add CO2 kind of
to their analysis and so wewere able to bring in CO2
into both their production facilities.
They use a lot of rail aswell as truck incorporate CO2
into the network and beable to provide to them
as kind of a almost free byproduct
of the model, kind of that CO2 analysis.
(23:51):
And they could run scenariosas well as they were looking
at different sites atdifferent sourcing strategies,
different customer kindof assignments, you know,
what the impact would be on CO2.
So that was an example wherekind of extending the scope
to other factors was one example there.
I'll give another examplewhere a medical drug company
(24:11):
is evaluating their productionlocations around the world.
And in this example, there'sobviously some tariffs
and duties, but also thereis some country rules
around local production andhow the government provides
and tenders and the winrates that you'll achieve.
So at the Milena's point aboutdemand, this was a very much
a situation where depending
on where the production location was,
(24:33):
the demand could shiftly change.
And so we helped evaluatenot only the network design
from a typical traditionalsense, but also how would
that location decision affect demand?
Which was significant in their example.
- And if the time cycleis speeding up as we work
on supply chain design, arethese iterative processes then?
(24:55):
Do they become almost aconstant in the background
or how would you describelike, you know, these sound
like some pretty massive projects.
They multi-year, multi-month?
Do you plan on a certain periodic rate?
- Yeah, so I think, you knowreally from what we're seeing
in Coupa's perspectiveand what we're seeing
in the environment is not only
(25:16):
are the strategic decisionshappen more frequently,
again, we've extended nowinto that tactical scope
and we see a repeatability requirement
that I would say 95% of ourcustomers are requiring,
and we certainly agree with that, right?
Instead of doing that episodicevent based kind of process,
it's now how do we make this repeatable so
(25:39):
that the data process can go from,
in many examples it's significant 30
to 60, 90 days down to hours, right,
to refresh a model.- Wow.
- And that maybe seem likea snik of improvement.
It is, but it's really not that hard
if we really just workthrough it, you know, one time
(25:59):
and then we get a compounding benefit.
'Cause now we can answerthose strategic questions
more frequently.
We can answer the tactical questions.
In many cases now we answera lot of additional questions
that we couldn't answer before.
So one example is cost toserve, we often get answers,
okay now I just needto know cost to serve.
Can you help me with that?
Well if we build this repeatable process,
(26:20):
we already have the models in place.
Cost to serve is a naturalbyproduct, we kind of get
for free again, right?
Can you answer now the
more tactical production planning problem?
Yes, because we canrefresh the data quickly
and we can provide those results.
And what we see in addition tothat is not only are we able
to solve these strategic andtactical problems frequently
(26:42):
because we have a setof data that's validated
and kind of refreshed and clean, we see
to your comments earlier,a lot of other parts
in the organizationwanna go access that data
and leverage just from a BI perspective
and data analyticsperspective, which we love
because now we're allworking on the same page
and to, from an organizationalbuy-in perspective,
(27:03):
now there's more incentivefor all the departments
and functions to providegood data to the system
because they wanna see theresults out of the systems
if you will, so we see thisvery much compounding benefit
if we can get thatrepeatability put in place.
So again that's a, I wouldalmost say a must have
in this day and age.
- So Milena, do youalso have some examples
(27:25):
from the work that you're doing?
- Yes, so one example isa pharmaceutical company
that we are currently working with and
that is directly related to this idea
of extending the scopeof supply chain design
and having a morecustomer-centric approach.
So in the pharma sector wehave a few industries trends
(27:47):
that are really redefining the way
that these companies are thinkingabout their supply chain.
And these are portfolioshifts towards different types
of drugs that are targetingmuch smaller audiences
but have potential much higher revenue.
Increase market competitionwith some generic competition
that could directlyimpact the market share
(28:08):
for their products.
And basically what they arestarting to realize is that
in order to increase their revenue
and market share, the drugis only one of the component
of the patient experience.
And so the way that you'llactually fulfill that drug
and bring it to yourpatient is a key element
that they want to consider going forward.
(28:31):
So first project that we workedwith them actually looked
at different last mile strategies
to deliver a certain drug to the patients
and they were exploring varioussupply chain configurations,
but also basically theresponse of the consumers
to things like home deliveryversus getting your drug
(28:53):
at a traditional channel,which would be a hospital
or pharmacy.
And we built a model thatwas basically incorporated
that customer response
with our overarchingsupply chain decision.
So that was an interesting exercise.
And in the second iterationof that, what we realized is
that in the specificspace there's actually
a much broader healthcare ecosystem
(29:14):
that we need to think about.
And so rather than justfocusing on a company
and a patient, we are now incorporating
additional stakeholderslike healthcare authorities,
insurance companies, et cetera, et cetera,
which also have a say in this space.
And so we're basically moving
towards a kind of a multi-actor model
that captures thoserelationships that exist
(29:35):
between those problems.
And as you see, it's highly customized
and highly specific to theoperations of this company.
And it's also going to be derived
into a series ofdifferent models according
to the country where you're operating
because of course the regulationsaround drug distribution
and such will be very different.
(29:55):
The second example that I have is linked
to this idea of incorporating the tactical
and the strategic.
And so there we are working with a company
that is a global shipping company.
We started with them by looking
at their overall distribution network.
So you know, I'm spanningfrom Asia many to the US and
(30:19):
as the time went by weactually realized that
in the current context weneeded to incorporate into
that strategic model a muchmore granular description
of the specific events thatcould arise, such as strikes
at ports or different weather conditions
that can delay shipments in certain areas.
And so now we are moving towards a model
(30:40):
that basically is incorporatingthe tactical scheduling
and planning and routeoptimization as well
as more strategic decisionson network design.
- I'm also curious, what areyou personally excited about?
Like what gets you upin the morning related
to these two questions?
(31:01):
Like you come to work andyou say, you know what,
I'm gonna help this companyovercome this obstacle.
What is that obstacle andwhy does it excite you?
- Yeah, so as far as you know,what keeps us excited here
at Coupa and me personally, Imean I've been doing this now
for 10 plus years at Coupa, LLamasoft.
And to your question, thereason I'm still here is
(31:22):
'cause I get more and I'mexcited 'cause I think
like we can make a differenceand we can help companies
and they can really find value.
We have the tools, thealgorithms, you know,
those algorithms are evolving,
but there are solid algorithmsthat we know how to use
and it's really kind of taken advantage
of the data that's not available,
the computational capabilities we have.
(31:43):
And then all of thatinfrastructure to kind of make
that a seamless process so
that we can answer questions quickly
for customers where beforethat took a long time to do.
Additionally, with cloudinfrastructure, you know,
we now have the ability toroll out apps and other things
that have kind of a user interface
that's much more customized to
(32:04):
that business function or that user.
But under the hood it's allthe math that we love and know.
All the algorithms and stuff are running
underneath the hood thatwe developed for them,
but now we're presentingit to more users in
that organization to take advantage of
that capability, run their own scenarios
within some guardrails typically,
(32:25):
but run their own scenarios,get their own results,
and allow the modelers to notbe still excited and be part
of that, but allow morepeople to take advantage
of these capabilities in the future.
- How do you incorporatedesigning for uncertainty
and anticipating the waythe discipline will evolve?
- So that's a really goodquestion because we are
(32:47):
at a phase where, as mentioned before,
we do not have this one model
or one framework fits everything.
And so we need to constantlybasically reinvent the way
that we are thinkingabout supply chain design.
And the first way that wewould do that would be not
(33:09):
so much about replacing the types of tools
that we are using, but really focusing
on the organizational processes,
the decision making processesaround supply chain design.
And here I think the key is really
to enable organizational learning in a way
that we can always adapt bothour designs and our processes.
And when we talk aboutorganizational learning,
(33:31):
there are several levels atwhich we can characterize it.
So the first and mostimmediate one is to say,
well we implement a design.
We observe a performance of that design.
For example, I know demand was higher,
capacity was exceeded,something like that.
And then we tweak our designto adapt to this new condition.
The second level is, Iwould say is more complex.
(33:53):
It's basically reframing the problem.
So we've mentioned some
of that, extending the scope, et cetera.
And so it's basically saying,I'm no longer focusing
on cost, but incorporating risk,
incorporating value creation,et cetera, et cetera.
And then the third level I would say
is even more challenging.
And it's about basically reflecting
on our decision makingprocess in the organization
(34:15):
and monitoring how thatprocess happens and trying
to improve that design process itself.
And that means who is involved
when we are defining our objectives?
Are we including the rightpeople in the organization?
The right people outsideof the organization?
Who's responsible formonitoring the results?
(34:36):
How are enabling these feedback loops?
And I say that mostcompanies currently may be
at the level one or leveltwo and are not still at
that level where theyare actually reflecting
on their process itself.
And that's one of the things
that I find actually the most rewarding.
It's when we get people fromfunctions in the organization
(34:59):
in the same room and we havethem come up with new problems,
with new ways of solving those problems
and understanding the valueof actually changing the way
that they are currently performing
their supply chain design.
- So what stops people from knowing
that they need to do that?
I mean, I believe the whitepaper probably is groundbreaking
(35:21):
in that respect.
People have never stood backand asked themselves like,
hey, I need to know, oh I've gotta talk
to people across my domains.
Or oh, I have access to this other data,
or oh, I need to do this moreperiodically than I have been.
What are the biggest obstacles?
'Cause we don't know whatwe don't know, right?
If we're running an organization
and we've been doing it this way.
- Yeah, so I think from what we see,
(35:44):
I would say there's probably three things
that get in the way of customerskind of moving forward.
The first relates to what I'llcall organizational buy-in.
Which from the leadershipperspective, you know,
we talked about earliercross-functional kind of involvement
and commitment.
IT as well as the analytical resources.
(36:04):
You know, we can say a CoE, acenter of excellence related
to kind of supply chain design.
All of those pieces need to be in place
and agreed upon such that
from an organizational perspective,
we are gonna sustainthis process over time.
And just from a CoE perspective,we know today obviously
with some of the challengesin the job market is, even
(36:25):
as CoE you have to have a process
of training the people,giving 'em more opportunities
to learn, recruitment, retention.
All of those things are critical
to keep the CoE kind ofprogressing and lively and growing.
The second thing is, I would say is
that companies do not have a clear roadmap
of the types of problems they wanna solve
and how they're gonna grow that over time.
(36:47):
I think the Milena's pointis, they're solving a problem
but they really don't know whatproblem two or problem three
and how those are connectedand what the sequence should be
and all of those things.
In addition to that projectroadmap we talked about earlier,
the idea that there needsto be an automated process
to kind of do the data collection
and the data foundationof this whole process.
(37:08):
So they lack kind of that infrastructure
to do the data collectionand data automation.
And so therefore everyprojects is a challenge.
And so people come to them and say,
"I need this question answered."
And the response is,
"Great, I'll have itbetween you in six months."
Well we know that's nolonger even acceptable.
So if you don't have thatcommon data infrastructure
in place, then you're notgonna be able to answer
(37:29):
to the questions timely andtherefore the value proposition
to the company is less.
The third thing is really thatwe have a sound understanding
in that organizationalbuy-in of the metrics
and the deliverables thatwe're gonna be providing
to the business.
So we know that there's an investment
in resources, an effort to do all this.
(37:50):
We need to be able to measureand show that to the business
that we are delivering the results.
I mean, historically Ithink we've been very good
at the analytics side,you know, maybe most
of the modeling team
and those data analytic peopleare not good salesmen, right?
So we need to helpsupport them to make sure
that hey, you are deliveringvalue to the business
and we actually provethat ROI to the business.
(38:10):
So those are the thingsthat I would say we see
as some of the common obstacles
to kind of really get him moving forward.
- So what steps should companies take then
to get the ball rolling?
- So there are many steps,
and I think Mike hasalready hinted too few
of them, such as, you know,increasing data availability,
making sure we haveautomated data processing
(38:32):
and sharing structuresin order to really get
that end-to-end view and transparency.
However, and here I'm seconding Mike
in what he previously saidis that I really do think
that the most important step is
at the level of organization.
And so we really need tohave a clear ownership
of the supply chain analytics and design.
(38:52):
For example, establish a dedicated team
or center of excellencewho's going to have
that long-term vision aroundsupply chain design and
also empowering that team bythe top levels of leadership.
So we wanna avoid to havethat be a separate cell
that has to battle with eachindividual structure function
in the organization.
And we want to have atop level sponsoring of
(39:16):
that supply chain design,central of excellence
or a dedicated team.
- We see things evolving very quickly
in many domains, includingsupply chain management.
Where do you think we're heading?
- What we're seeing asfar as where we're heading
in the next several yearsis really several things.
And we touched upon some of those
throughout this discussion.
(39:37):
I mean, first of courseis we continue to see
that scale data, big data,the scale of the data,
the granularity of thedata that we need as well
as the cloud-based connected solutions
will continue to increase.
We already talked about the idea
that the strategictactical/planning solutions
are kind of continuing to merge
(39:58):
as far as what those problemswe're trying to solve.
I think as Milena noted though,
that doesn't mean thatthere's one size fits all,
but there's a library of solutions
that are kind of interconnectedas far as their data,
but there are maybe slightlydifferent variations
to solve different problems.
Kinda related to all thatis, you know, the persona
of continuous processdriven design, right?
That this is a continuous process
(40:20):
and we talked about the idea of
that requires then this repeatablekind of foundation to it.
We talked a little bit earlier
about this multienterprise kind of solution
that's gonna kind of grow.
So we do think that multienterprise kind of solutions
will continue to be playa part in the future here.
Risk of course, we don'tthink that's gonna go away.
We think risk will continueto be, play a prominent role
(40:42):
in kind of these network designsand how to be able to react
and respond quickly to those
as well as proactivelyplanned for those things.
Sustainability, of course,those kind of pieces
to the puzzle will continueto, I think be important.
And again, we talked a little bit earlier
as well about the AI machine learning.
We see again extending the solutions
(41:03):
to include more AI machinelearning components
to these problems, whetherit's kind of on data analysis
and trends or other components as well
as prescriptive kind of findings
and solutions that they might provide.
- So I think Mike coveredquite a broad range of things
that we are expecting tohappen over the next few years.
(41:24):
One thing that I wouldwant to add is relevant
to what we are observing now.
I would say that the lastfew years really triggered
a change in a way thatcompanies are thinking
about supply chain design.
And I think a lot of companiesare starting to be aware
of the different problems,starting to realize the necessity
(41:46):
to be more customer centric,to incorporate risk.
And we are at a stage wherethe problems are there
and well defined,
but we don't necessarilyhave the solution.
And I would say that supply chains are
in a very specific place.
They're really now at the center
of the corporate strategyand decision making.
There's a lot of excitement going on
(42:07):
and I see a lot of experimentationhappening, you know,
particularly with new products
and services, how do wedifferentiate from our competitors?
How do we use supplychain design to do that?
I do oversee that this experimentation is
not necessarily always supported
by this data-driven approach and
that the level of decisionmaking maturity is often limited.
(42:28):
And so I'm expecting that
in next few years we'llkind of see more clearly
what are some of the winningstrategies among those range
of different things thatpeople are trying out.
And basically identifiedkind of the winners
and losers of this process.
And there we will thinkthat having an intentional,
well-structured andanalytics driven approach
is gonna be key to being
in the winning side of the equation.
(42:51):
- Excellent. So how dopeople get the ball rolling?
If I'm an operator
in a company, how do Iget the ball rolling?
- Yeah, so obviously wework with a lot of companies
that are just getting started and
so what we typically wanna tell them to do
or how we coach them throughthat process is pick something
that's small but important to the business
so you can kind of work on something
(43:11):
that clearly is gonnademonstrate value to the business
and kind of show some of that value
as far as the effort that we do.
We certainly prefer somethingthat has some repeatability,
like we wanna get answers,you know, more frequently
and we wanna then build on top of
that kind of a repeatable process.
So let's get in place something
(43:32):
that not only answersquestions, but answers it
in a repeatable way so customers can begin
to see kind of how that, thevalue of that repeatability
and how it can speed a lot of the speed
to answer a lot of the questions.
Certainly there's manyorganizations out there, you know,
CTL, Coupa, other companiesthat can provide some support
to those companies to helpthem guide them along the way.
(43:54):
We certainly haveexperience with customers
that struggle maybe a littlebit more than they should.
An external kind of partnercan help kind us speed
that, you know, time into delivery.
I think overall being moreefficient for everyone.
And then of course,
leverage the existing solutions out there.
There's plenty of solutions out there
that already can kindof solve these problems.
(44:14):
Take advantage of what's outthere and make sure you fit it
to what you need, butyou know, find something
that's the right fit for youruse case and and get going.
- So Mike, Milena, thank youvery much for being here today.
- Thank you.
- Thank you. My pleasure.
Glad to participate.
- We really appreciate yourtime and super valuable to us
and hopefully it'll bevaluable to everyone.
(44:36):
So yeah, thank you.
- Thank you. Yeah, awesome.
- Thank you so much.
(upbeat music)
- All right everyone,thank you for listening.
I hope you've enjoyed this edition
of, "MIT Supply Chain Frontiers."
My name is Arthur Grau,Communications Officer
for the center, and I invite you
to visit us anytime atctl.mit.edu or search
(44:56):
for, "MIT Supply Chain Frontiers,"
on your favorite listening platform.
Until next time.
(upbeat music)
(upbeat music)
- Welcome to MIT Supply Chain Frontiers
from the MIT Center forTransportation and Logistics.
Each episode featurescenter researchers and staff
or experts from the fieldfor in-depth conversations
about business education and beyond.
(upbeat music)
(00:22):
- Thank you for joining today's,
"MIT CTL Supply Chain Frontiers," podcast.
So happy to have MikeBucci and Milena Janjevic.
Milena is a researchscientist here at MIT CTL.
And Mike, can you tell usa little bit about yourself
and your role?
- Yeah, sure thing.
Mike Bucci, I work forCoupa, formerly LLamasoft
(00:43):
as a few years back.
Been largely in the services,professional services group
during that time.
Mostly working with companies
around the world in all industries.
Kind of helped them understand how
to leverage the Coupa Software
to solve their supplychain design problems.
So I've been fortunate towork with many companies
(01:04):
around the globe and looking forward
to sharing some of thatexperiences here today.
- Excellent.
And Milena, can you tell usa little bit about your work
and how it relates to supply chain design
and the work that Mike is doing?
- Yes, thanks for having me here.
So I am leading the supplychain design initiative,
which is basically looking at the ways
in which we can improvethe decision making process
(01:28):
around supply chain design.
So here at MIT CTL, we do a variety
of industry sponsored researchand educational programs
with the aim of helpingcompanies learn how
to better address theirsupply chain design problems.
- Great, great, great.
So for today we're gonna betalking a little bit about sort
(01:49):
of old world or past use supplychain design and what some
of your research and somewhat new practices are
that you're doing in asupply chain design space.
So you recently co-publisheda white paper called,
The New Competitive Edge Analytics
Driven Supply Chain Design.
And in the white paper you mentioned
(02:10):
that most companies are still
using standard supplychain design practices
that have been in placefor around 30 years.
What is the traditionalsupply chain design approach?
- From the Coupa's side, Imean really what we've seen
over the years and therehas been a transition,
but yeah, the traditionalapproach and practice here
for supply chain design has really been
(02:31):
around more event-based orepisodic kind of processes
that occur at some schedule,
but usually less frequently every year
through two to three years.
And the idea behind that is really
to reassess a significancesupply chain problem,
strategic supply chain problem.
It could be a DC network design problem.
(02:52):
It could be productioncapacity planning problem,
but it's more of thatevent-based project or process.
And there's really a verysignificant heavy lift
to go find the data, collectthe data, cleanse the data.
You know, get organizationalbuy-in and the process.
Provide recommendations andthen go implement the solution.
(03:14):
And so that's kind of aprocess that reoccurs,
but it's kind of again, thismore of event-based problem.
And the focus there istypically on minimizing cost.
It includes, you know,obviously balancing service
and inventory, but it'sjust that general problem.
But again, it's more ofthat event-based activity.
- Does the periodicity or the time in
(03:36):
between updates affect thedifficulty of you to get the data
in a situation like that orto get accurate information?
- So I'd say that when you are redesigning
your supply chains every fiveyears or 10 years, the level
of granularity that youwill be able to incorporate
in those supply chain studies,it's typically very low.
(03:58):
So typically companies will aggregate data
at a very high level and that will lead
to all kinds of approximationsthat will ultimately result
in a big gap betweenthe expected performance
of your supply chain
and then the realized performance
once you implement a new design.
(04:19):
And some of that has been, Iwould say has resulted because
of simply the computational power
that was limited in the past.
But now we see that the data availability
has significantly increased,
but the supply chain design practices
have not necessarily caught up with that.
(04:40):
- Right. Anything elseto add on that point?
Like as far as the length of time between?
- Yeah, yeah, I think as Milena said,
I mean historically we'vebeen somewhat constrained
by computational power
and other things arecertainly looking forward
to what we see in the future.
What we see coming intoplay now is kind of
(05:00):
that more frequent analysisstudy which brings a whole bunch
of pros and cons related to that.
But we can talk through that.
- Yeah, well I mean let's talk about that.
So before we get to the pros,let's imagine that it's 1995
and every five years we'redoing a supply chain design.
The cons to that that I heard are
(05:20):
that you're just, you'regonna have highest level data,
you're not gonna have any granularity.
Are there any other consto that length in between?
- So one big industry trend
that we see is basically amuch higher pace of change.
And that's gonna beboth on the demand side
and on the supply side.
(05:43):
Companies are facingincreasing customer demands
and the need to readapt the supply chains
in much more frequentmatter to basically cater
to those increasing demands.
And that requires tohave a much higher speed
of adaptation of their
supply chain designs.
(06:04):
- Yeah, I mean thinking back,
as Milena said, I mean if we go
back decades here, thecomputational complexity
or the challenges we had forcomputational power, horsepower
as well as data collection, right?
Because systems were dispersed,uncentralized, unstructured
in many ways to try toconsolidate the data
and effectively leverage itfor these kind of studies,
(06:26):
that was largely the challenges.
So today we obviously havea significant difference,
right, where as Milena said,
supply chains are much morediverse, both from a supply
and demand perspective,there's more continual change
and the frequency of adaptinga supply chain is required
to be much more frequent than in the past.
(06:48):
And all that requires, of course
than the infrastructure around that.
So I think the newer landscape is
that we not only have acomputational capability, we
also have the systeminfrastructures to be able
to pull the data morefrequently and more efficiently
to kind of do thesestudies more frequently
(07:08):
and therefore leverage thoseresults obviously and react
to them much more quickerthan we could in the past.
- Great.
And this is gonna lead
to the conversation aboutsupply chain design.
Just before that, youmentioned consumer demand.
So people want things now.
They want things in theformat, they want 'em
on their front door or theywant 'em in a box somewhere
(07:28):
or maybe they want 'em delivered
to their summer home or whatever it is.
Is there a B2B demand that's changing?
I think is one question that I have.
And then what are some of the other trends
that are pushing towardsthis more tight timeframe
and advanced supply chain design models?
- So I would say thatcustomer centricity has become
a key element and a key sourceof competitive advantage
(07:51):
and both into B2C and B2B realm.
And that's, as Mike mentioned,that's not something
that is typically covered
by the traditional supplychain design methods,
which are really focusedon this cost minimization.
And so whether it's in theB2C or B2B space, we see
that delivery lead time
and delivery reliability arebecoming key order winners
(08:15):
and cannot be ignored fromsupply chain design studies.
- And from the Coupa side,
are you seeing peoplechanging their sourcing models
or is there anythinghappening with globalization
or does that also influence this process?
- Yeah, I would say absolutely.
I mean, we all know that supplychains have been diversified
(08:35):
and global now in nature.
And as we've seen obviously inthe pandemic and other things
along the last severalyears, one disruption along
that supply chain canhave a significant impact
on your entire system.
And so all of thesequestions now that were maybe
in the past more long-term questions
(08:57):
that we could answer once and not have
to revisit them frequentlynow they are questions
that we need to revisit often because
of all the things Milenadescribed, customer behavior
as well as on the supplychain, all the changes
that are occurring on the supply side.
I would also say that the need
to reassess these strategic decisions
more frequently leads kind of a shrinkage
(09:20):
between the strategic andtactical level decisions
that are occurring in companies.
And so now you're makingstrategic decisions
more frequently, maybe quarterly,
and you're making tacticaldecisions maybe monthly
or quarterly.
And now there's kind of a connection there
that we were trying to capture.
And that's one thing we seea lot with the companies
that we're working with is
(09:40):
that though there's a sniff overlap
between those kinds of questions
and those time horizons were typically
were separate kind ofquestions that were asked.
- I think we're understandingthe challenge now.
We know where the market is sitting,
we know where business players are sitting
and what they're up against.
And so in the white paper you mentioned
that there are four opportunities,primary opportunities
(10:01):
that companies and and managers can take
to reimagine the designof their supply chains.
And they are extending the scope,
incorporating the tactical.
Accounting for risk andplanning for resilience
and adopting new technologiesand business models.
So I'd love to get a coupleminutes in on each of these
(10:22):
to find out more about what they mean
and maybe what they mean on the ground
and what you've beenseeing with the people
that you've been working with.
So let's start with extending the scope.
What do you mean by extending the scope?
- Yeah, so the first andmost immediate opportunity
that we see here isextending the objectives
that we are considering inour supply chain design.
(10:44):
So we've already mentioned thattraditional methods focused
on cost minimization on physical structure
of a supply chain.
And that's a very limiting wayof considering supply chains
and supply chain design inthe contemporary market.
And so supply chain designshould have a much broader scope,
(11:05):
which basically focuses onlong-term value creation
and basically consider theinteractions that happen
between the choices we makein the supply chain design
and our ability to generatevalue, increase market share,
and generate revenue.
So for example, one keyquestion in supply chain design
(11:27):
that we have been answering
for the past 30 yearsis where should I look
at my facilities?
How many of thosefacilities should I have?
And if you have purely acost based perspective,
you know the answer's probablygoing to be to centralize
as much as you can andto gain some economies
of scale, have centralizedinventory with lower cost.
(11:49):
But if you consider thisidea of customer centricity
and the fact that the market share
that you can capture willdepend on your ability
to actually serve your clients
in a reliable and fast fashion,then the answer is going
to be completely different.
I do not see that a lotof companies are currently
(12:09):
using approaches that fully account for
that aspect of value creation.
- Yeah, I think Milena covered a lot
of the items I would mention.
I would also say that, you know,
as far as extending thescope, I mean historically
as Milena kind ofoutline, if we're looking
at a supply chain problem,if we're for example looking
at a DC a distribution networkkind of study, we'll put
(12:33):
in the distribution network, some
of the customers maybe alittle bit on the sourcing,
but now we have the abilityto go further back, right?
Maybe we can pick up somemore supplier detail.
We may even include options
of supplier direct kindof shipments, other ways
that we can kind of extend the breadth
of the scope of the supply chain problem.
(12:53):
That's one thing.
And then of course addingin other components or costs
or considerations to the model as well.
Milena obviously discussedthe idea of being closer
to the customer, the impact onhow that can have on demand.
We of course can incorporatethings like taxes and duties
and other kind of new governmentregulations as well as far
as understanding how oursupply chain design should be,
(13:16):
sustainability is anothercomponent as well, right?
Companies are moreconcerned with their CO2
or total emissions and other components.
We can incorporate that aswell kind of into the problem
and evaluate that in conjunctionwith cost, et cetera.
- So that's a little bitabout extending the scope
and broadening what youtraditionally would consider
(13:36):
as a supply chain design problem, right?
So tell us a little bitmore about what you mean
by incorporating the tactical?
- Yeah, I'll start here.
I mean, I think incorporatingthe tactical really is again,
this combined or shrinkingof the difference
between the strategicand tactical decisions.
And in all the work thatwe do with our clients
and customers that we workwith, we see when we try
(13:58):
to address both the strategic question
and the tactical questions,you know, we estimate
that 75% of the datarequirements are the same.
And so why not take advantageand get a multiplier of
that effort to solve bothkinds of problems and have
that same kind of data foundation
where there's not discrepancies
between the tacticaldecision making process
(14:19):
and the strategic process.
So we see a lot of convergence there
of incorporating tactical level detail
into our strategic models.
We're kind of overlapping
the two kind of decision making processes.
And examples of thatwould be kind of things
like an S&OP process,short-term, rough cut,
capacity planning,replenishment inventory planning
(14:42):
kind of policies.
You know, how can we address those more
in our supply chain designkind of problem optimizations.
- One thing that I wouldadd to that is that
from the modeling perspective
and data collectionperspective, the incorporation
of these tactical decisions will
often require a certain effort.
(15:04):
And the idea is not
that every company shouldincorporate each type
of tactical decision with thesame level of granularity,
but really identify those area
that are key for their value creation.
So if I am competing in thelast mile space, I probably need
(15:27):
to have a much more granularand precise integration
of my routing decisionsand my inventory decisions
in the last mile thanif I am a manufacturer
that is mainly going to be looking
at production planning decisions
as a key driver of their advantage.
And so this idea that wecan have a single model
(15:50):
and a single approach
that would fit alldifferent industry context
and companies is now outdated.
- Excellent. I like that, I like that.
So not only is it tactical,
but the tacticalness isgonna depend on where you are
in the supply chain, what your business is
and that sort of thing.
(16:11):
So the next of the fouropportunities that you recognize
in the white paper is to account for risk
and plan for resilience.
And I think this is on everybody's mind
after the last few years.
So can you tell us moreabout accounting for risk
and planning for resilience?
- So yeah, so I think the firststep here is for companies
to basically map out the different sources
(16:35):
of risk, differentsources of vulnerability
and to characterize those.
And there we need to recognize
that when we talk about risk,there are different categories
of risk that will be accounted for
in a very different mannerin our supply chain design.
So if I'm talking about,you know, natural demand
(16:57):
that just comes from somevariations in the markets,
that's very different than a risk linked
to a major natural disaster or pandemic.
And so the way I will account for that
in my supply chain design and
in my tools should bevery, very different.
And so I think that
(17:18):
at this point companiesare very much aware
of the requirement toincorporate the risk,
but they don't necessarilyhave the right tools to employ
and address each type ofrisk that they might face.
- Yeah, to add to what Milena said,
I mean what we're seeingis historically companies
(17:40):
will run scenarios fortheir supply chain design
or tactical problems and they'lldo a few sensitivities of,
well what if TransportationCostco up or down 5%.
But we're seeing obviouslythose sensitivities
increase significantly atwo to three x increase
in number of scenarioscompanies are running today.
Partially due to thisrisk related problem.
(18:03):
So the first thing we see isjust running more scenarios
to look at the resiliency of your network
to different changes.
They can be maybe their cost changes
or they could be changes to constraints.
I have a certain port volume
that I'm expecting to move through.
What if that port volumewere to decrease by 50%?
What would my supply chain,how would my supply chain react
(18:25):
and what is the impactof that on my network?
A second theory to that aswell is we're now able to pull
into some of these models risk scores
and risk metrics, you know,from external sources.
And so by pulling that in someways we can kind of put that
into the objectivefunction or we're trying
to minimize certain amountof risks in our supply chain,
(18:46):
or at least report that out as far
as certain nodes, maybe we'reflowing 95% of our volume
through a certain node.
We wanna decide to highlightthat is a potential risk
because we were kind of single sourced
or single node kind of constrained
if there was a risk eventrelated to that node.
Also, again, based on thoseexternal risk metrics,
(19:10):
we can provide some risk scorefor our supply chain design,
you know, based upon a supplier's risk,
based on differenttransportation risks, et cetera.
And so we're seeing that more and more
in the kind of themodels that we're running
and the requirements our customers
are having regarding risk and resilience.
- I think what's fascinating about this
then is it kind of leads us directly
(19:31):
into the fourth opportunitybecause I'm sensing
that people are maybe collaborating more
or maybe they're gettingdata from resources
that didn't exist beforethat are public or
that are, you know, able to be purchased
and maybe there are new ways
of analyzing and new technology.
So can you tell us aboutthe fourth opportunity
to adopt new technologiesand business models?
(19:53):
- Yeah, I'll start here.
So for what we're seeing,
and you're right on there,Arthur, I mean as far
as what we're seeing is, again,risk is one good example.
We're able to maybeincorporate external data more,
consider that, or atleast, you know, define
that in our solutions.
I would say a couple things here in play,
of course as you mentioned,it's this big data, right?
(20:14):
There's a lot more data we can capture
and bring into our analysis.
The cloud infrastructure iscertainly helping us there
as well, right?
With more information around the cloud,
we can do some interconnectivity.
It's easier to get datathan it was in the past
from different sourcesand different systems
within a company and externally.
AI machine learning playinto that as well, right?
(20:36):
Now we have models wherewe're doing the optimization
or the simulation, butmaybe on the front end
and the back end of thatwe're doing machine learning
and AI to kind of tease outother things in the solution
to better provide results tothe customer, which some of
that can be related to thatrisk and resiliency as well.
And I even think furtherwe're seeing situations where
(20:57):
that AI machine learningcan actually look for things
in our supply chain
that may be I'll saycognitive blind spots maybe
to the modeler, but wedon't really see that issue
or we don't see that problem.
But machine learning orAI can kind of pick up
on potential issues with our supply chain.
It could be risk related orit could be kind of trends
(21:19):
and costs and other thingsthat we're seeing over time
in their networks.
So those are some examples
where I think the newtechnologies are kind of morphing
into this kind of supplychain design problem.
- I would add one thing to that is
that the digital transformationthat Mike has just mentioned
(21:39):
also enables a new types oforganizational relationships
between differentcompanies and relationships
that now include actors thatwere previously not part
of our supply chain ecosystem.
One example of that is the use
of crowdsourced resources,it being transportation
(22:01):
or warehousing and basicallyeverything that is targeted
towards an on-demand useof the resources rather
than investment in assets,
which is a really interestingmitigation strategy
that can build higher resilienceinto your supply chain.
When we talked about thetraditional approaches
(22:22):
to the supply chain designa day would typically assume
a very classical supplierbuyer relationship
that did not reflect these
new complex organizationalrelationships, did not,
for example, incorporate the ways
in which we would sharerevenue, share costs
and share risks amongthose different parties.
(22:44):
So that's another opportunity that we see.
It's extending basically the way
that we represent theorganizational structure
of the supply chains to take advantage
of these new business models
that have arised in the last few years.
- So we've seen the four opportunities are
to extend the scope ofthe supply chain design
(23:06):
to incorporate tactical, moregranular information and data.
Accounting for risk andadopting new technologies.
Do you have any examples of many of these
that you wanna sharefrom your recent work?
- One would be a miningcompany that we've worked with
for many years and they're doing a lot
of strategic analysis,even some tactical studies
(23:28):
on their capacity planning.
But recently they wantedto add CO2 kind of
to their analysis and so wewere able to bring in CO2
into both their production facilities.
They use a lot of rail aswell as truck incorporate CO2
into the network and beable to provide to them
as kind of a almost free byproduct
of the model, kind of that CO2 analysis.
(23:51):
And they could run scenariosas well as they were looking
at different sites atdifferent sourcing strategies,
different customer kindof assignments, you know,
what the impact would be on CO2.
So that was an example wherekind of extending the scope
to other factors was one example there.
I'll give another examplewhere a medical drug company
(24:11):
is evaluating their productionlocations around the world.
And in this example, there'sobviously some tariffs
and duties, but also thereis some country rules
around local production andhow the government provides
and tenders and the winrates that you'll achieve.
So at the Milena's point aboutdemand, this was a very much
a situation where depending
on where the production location was,
(24:33):
the demand could shiftly change.
And so we helped evaluatenot only the network design
from a typical traditionalsense, but also how would
that location decision affect demand?
Which was significant in their example.
- And if the time cycleis speeding up as we work
on supply chain design, arethese iterative processes then?
(24:55):
Do they become almost aconstant in the background
or how would you describelike, you know, these sound
like some pretty massive projects.
They multi-year, multi-month?
Do you plan on a certain periodic rate?
- Yeah, so I think, you knowreally from what we're seeing
in Coupa's perspectiveand what we're seeing
in the environment is not only
(25:16):
are the strategic decisionshappen more frequently,
again, we've extended nowinto that tactical scope
and we see a repeatability requirement
that I would say 95% of ourcustomers are requiring,
and we certainly agree with that, right?
Instead of doing that episodicevent based kind of process,
it's now how do we make this repeatable so
(25:39):
that the data process can go from,
in many examples it's significant 30
to 60, 90 days down to hours, right,
to refresh a model.- Wow.
- And that maybe seem likea snik of improvement.
It is, but it's really not that hard
if we really just workthrough it, you know, one time
(25:59):
and then we get a compounding benefit.
'Cause now we can answerthose strategic questions
more frequently.
We can answer the tactical questions.
In many cases now we answera lot of additional questions
that we couldn't answer before.
So one example is cost toserve, we often get answers,
okay now I just needto know cost to serve.
Can you help me with that?
Well if we build this repeatable process,
(26:20):
we already have the models in place.
Cost to serve is a naturalbyproduct, we kind of get
for free again, right?
Can you answer now the
more tactical production planning problem?
Yes, because we canrefresh the data quickly
and we can provide those results.
And what we see in addition tothat is not only are we able
to solve these strategic andtactical problems frequently
(26:42):
because we have a setof data that's validated
and kind of refreshed and clean, we see
to your comments earlier,a lot of other parts
in the organizationwanna go access that data
and leverage just from a BI perspective
and data analyticsperspective, which we love
because now we're allworking on the same page
and to, from an organizationalbuy-in perspective,
(27:03):
now there's more incentivefor all the departments
and functions to providegood data to the system
because they wanna see theresults out of the systems
if you will, so we see thisvery much compounding benefit
if we can get thatrepeatability put in place.
So again that's a, I wouldalmost say a must have
in this day and age.
- So Milena, do youalso have some examples
(27:25):
from the work that you're doing?
- Yes, so one example isa pharmaceutical company
that we are currently working with and
that is directly related to this idea
of extending the scopeof supply chain design
and having a morecustomer-centric approach.
So in the pharma sector wehave a few industries trends
(27:47):
that are really redefining the way
that these companies are thinkingabout their supply chain.
And these are portfolioshifts towards different types
of drugs that are targetingmuch smaller audiences
but have potential much higher revenue.
Increase market competitionwith some generic competition
that could directlyimpact the market share
(28:08):
for their products.
And basically what they arestarting to realize is that
in order to increase their revenue
and market share, the drugis only one of the component
of the patient experience.
And so the way that you'llactually fulfill that drug
and bring it to yourpatient is a key element
that they want to consider going forward.
(28:31):
So first project that we workedwith them actually looked
at different last mile strategies
to deliver a certain drug to the patients
and they were exploring varioussupply chain configurations,
but also basically theresponse of the consumers
to things like home deliveryversus getting your drug
(28:53):
at a traditional channel,which would be a hospital
or pharmacy.
And we built a model thatwas basically incorporated
that customer response
with our overarchingsupply chain decision.
So that was an interesting exercise.
And in the second iterationof that, what we realized is
that in the specificspace there's actually
a much broader healthcare ecosystem
(29:14):
that we need to think about.
And so rather than justfocusing on a company
and a patient, we are now incorporating
additional stakeholderslike healthcare authorities,
insurance companies, et cetera, et cetera,
which also have a say in this space.
And so we're basically moving
towards a kind of a multi-actor model
that captures thoserelationships that exist
(29:35):
between those problems.
And as you see, it's highly customized
and highly specific to theoperations of this company.
And it's also going to be derived
into a series ofdifferent models according
to the country where you're operating
because of course the regulationsaround drug distribution
and such will be very different.
(29:55):
The second example that I have is linked
to this idea of incorporating the tactical
and the strategic.
And so there we are working with a company
that is a global shipping company.
We started with them by looking
at their overall distribution network.
So you know, I'm spanningfrom Asia many to the US and
(30:19):
as the time went by weactually realized that
in the current context weneeded to incorporate into
that strategic model a muchmore granular description
of the specific events thatcould arise, such as strikes
at ports or different weather conditions
that can delay shipments in certain areas.
And so now we are moving towards a model
(30:40):
that basically is incorporatingthe tactical scheduling
and planning and routeoptimization as well
as more strategic decisionson network design.
- I'm also curious, what areyou personally excited about?
Like what gets you upin the morning related
to these two questions?
(31:01):
Like you come to work andyou say, you know what,
I'm gonna help this companyovercome this obstacle.
What is that obstacle andwhy does it excite you?
- Yeah, so as far as you know,what keeps us excited here
at Coupa and me personally, Imean I've been doing this now
for 10 plus years at Coupa, LLamasoft.
And to your question, thereason I'm still here is
(31:22):
'cause I get more and I'mexcited 'cause I think
like we can make a differenceand we can help companies
and they can really find value.
We have the tools, thealgorithms, you know,
those algorithms are evolving,
but there are solid algorithmsthat we know how to use
and it's really kind of taken advantage
of the data that's not available,
the computational capabilities we have.
(31:43):
And then all of thatinfrastructure to kind of make
that a seamless process so
that we can answer questions quickly
for customers where beforethat took a long time to do.
Additionally, with cloudinfrastructure, you know,
we now have the ability toroll out apps and other things
that have kind of a user interface
that's much more customized to
(32:04):
that business function or that user.
But under the hood it's allthe math that we love and know.
All the algorithms and stuff are running
underneath the hood thatwe developed for them,
but now we're presentingit to more users in
that organization to take advantage of
that capability, run their own scenarios
within some guardrails typically,
(32:25):
but run their own scenarios,get their own results,
and allow the modelers to notbe still excited and be part
of that, but allow morepeople to take advantage
of these capabilities in the future.
- How do you incorporatedesigning for uncertainty
and anticipating the waythe discipline will evolve?
- So that's a really goodquestion because we are
(32:47):
at a phase where, as mentioned before,
we do not have this one model
or one framework fits everything.
And so we need to constantlybasically reinvent the way
that we are thinkingabout supply chain design.
And the first way that wewould do that would be not
(33:09):
so much about replacing the types of tools
that we are using, but really focusing
on the organizational processes,
the decision making processesaround supply chain design.
And here I think the key is really
to enable organizational learning in a way
that we can always adapt bothour designs and our processes.
And when we talk aboutorganizational learning,
(33:31):
there are several levels atwhich we can characterize it.
So the first and mostimmediate one is to say,
well we implement a design.
We observe a performance of that design.
For example, I know demand was higher,
capacity was exceeded,something like that.
And then we tweak our designto adapt to this new condition.
The second level is, Iwould say is more complex.
(33:53):
It's basically reframing the problem.
So we've mentioned some
of that, extending the scope, et cetera.
And so it's basically saying,I'm no longer focusing
on cost, but incorporating risk,
incorporating value creation,et cetera, et cetera.
And then the third level I would say
is even more challenging.
And it's about basically reflecting
on our decision makingprocess in the organization
(34:15):
and monitoring how thatprocess happens and trying
to improve that design process itself.
And that means who is involved
when we are defining our objectives?
Are we including the rightpeople in the organization?
The right people outsideof the organization?
Who's responsible formonitoring the results?
(34:36):
How are enabling these feedback loops?
And I say that mostcompanies currently may be
at the level one or leveltwo and are not still at
that level where theyare actually reflecting
on their process itself.
And that's one of the things
that I find actually the most rewarding.
It's when we get people fromfunctions in the organization
(34:59):
in the same room and we havethem come up with new problems,
with new ways of solving those problems
and understanding the valueof actually changing the way
that they are currently performing
their supply chain design.
- So what stops people from knowing
that they need to do that?
I mean, I believe the whitepaper probably is groundbreaking
(35:21):
in that respect.
People have never stood backand asked themselves like,
hey, I need to know, oh I've gotta talk
to people across my domains.
Or oh, I have access to this other data,
or oh, I need to do this moreperiodically than I have been.
What are the biggest obstacles?
'Cause we don't know whatwe don't know, right?
If we're running an organization
and we've been doing it this way.
- Yeah, so I think from what we see,
(35:44):
I would say there's probably three things
that get in the way of customerskind of moving forward.
The first relates to what I'llcall organizational buy-in.
Which from the leadershipperspective, you know,
we talked about earliercross-functional kind of involvement
and commitment.
IT as well as the analytical resources.
(36:04):
You know, we can say a CoE, acenter of excellence related
to kind of supply chain design.
All of those pieces need to be in place
and agreed upon such that
from an organizational perspective,
we are gonna sustainthis process over time.
And just from a CoE perspective,we know today obviously
with some of the challengesin the job market is, even
(36:25):
as CoE you have to have a process
of training the people,giving 'em more opportunities
to learn, recruitment, retention.
All of those things are critical
to keep the CoE kind ofprogressing and lively and growing.
The second thing is, I would say is
that companies do not have a clear roadmap
of the types of problems they wanna solve
and how they're gonna grow that over time.
(36:47):
I think the Milena's pointis, they're solving a problem
but they really don't know whatproblem two or problem three
and how those are connectedand what the sequence should be
and all of those things.
In addition to that projectroadmap we talked about earlier,
the idea that there needsto be an automated process
to kind of do the data collection
and the data foundationof this whole process.
(37:08):
So they lack kind of that infrastructure
to do the data collectionand data automation.
And so therefore everyprojects is a challenge.
And so people come to them and say,
"I need this question answered."
And the response is,
"Great, I'll have itbetween you in six months."
Well we know that's nolonger even acceptable.
So if you don't have thatcommon data infrastructure
in place, then you're notgonna be able to answer
(37:29):
to the questions timely andtherefore the value proposition
to the company is less.
The third thing is really thatwe have a sound understanding
in that organizationalbuy-in of the metrics
and the deliverables thatwe're gonna be providing
to the business.
So we know that there's an investment
in resources, an effort to do all this.
(37:50):
We need to be able to measureand show that to the business
that we are delivering the results.
I mean, historically Ithink we've been very good
at the analytics side,you know, maybe most
of the modeling team
and those data analytic peopleare not good salesmen, right?
So we need to helpsupport them to make sure
that hey, you are deliveringvalue to the business
and we actually provethat ROI to the business.
(38:10):
So those are the thingsthat I would say we see
as some of the common obstacles
to kind of really get him moving forward.
- So what steps should companies take then
to get the ball rolling?
- So there are many steps,
and I think Mike hasalready hinted too few
of them, such as, you know,increasing data availability,
making sure we haveautomated data processing
(38:32):
and sharing structuresin order to really get
that end-to-end view and transparency.
However, and here I'm seconding Mike
in what he previously saidis that I really do think
that the most important step is
at the level of organization.
And so we really need tohave a clear ownership
of the supply chain analytics and design.
(38:52):
For example, establish a dedicated team
or center of excellencewho's going to have
that long-term vision aroundsupply chain design and
also empowering that team bythe top levels of leadership.
So we wanna avoid to havethat be a separate cell
that has to battle with eachindividual structure function
in the organization.
And we want to have atop level sponsoring of
(39:16):
that supply chain design,central of excellence
or a dedicated team.
- We see things evolving very quickly
in many domains, includingsupply chain management.
Where do you think we're heading?
- What we're seeing asfar as where we're heading
in the next several yearsis really several things.
And we touched upon some of those
throughout this discussion.
(39:37):
I mean, first of courseis we continue to see
that scale data, big data,the scale of the data,
the granularity of thedata that we need as well
as the cloud-based connected solutions
will continue to increase.
We already talked about the idea
that the strategictactical/planning solutions
are kind of continuing to merge
(39:58):
as far as what those problemswe're trying to solve.
I think as Milena noted though,
that doesn't mean thatthere's one size fits all,
but there's a library of solutions
that are kind of interconnectedas far as their data,
but there are maybe slightlydifferent variations
to solve different problems.
Kinda related to all thatis, you know, the persona
of continuous processdriven design, right?
That this is a continuous process
(40:20):
and we talked about the idea of
that requires then this repeatablekind of foundation to it.
We talked a little bit earlier
about this multienterprise kind of solution
that's gonna kind of grow.
So we do think that multienterprise kind of solutions
will continue to be playa part in the future here.
Risk of course, we don'tthink that's gonna go away.
We think risk will continueto be, play a prominent role
(40:42):
in kind of these network designsand how to be able to react
and respond quickly to those
as well as proactivelyplanned for those things.
Sustainability, of course,those kind of pieces
to the puzzle will continueto, I think be important.
And again, we talked a little bit earlier
as well about the AI machine learning.
We see again extending the solutions
(41:03):
to include more AI machinelearning components
to these problems, whetherit's kind of on data analysis
and trends or other components as well
as prescriptive kind of findings
and solutions that they might provide.
- So I think Mike coveredquite a broad range of things
that we are expecting tohappen over the next few years.
(41:24):
One thing that I wouldwant to add is relevant
to what we are observing now.
I would say that the lastfew years really triggered
a change in a way thatcompanies are thinking
about supply chain design.
And I think a lot of companiesare starting to be aware
of the different problems,starting to realize the necessity
(41:46):
to be more customer centric,to incorporate risk.
And we are at a stage wherethe problems are there
and well defined,
but we don't necessarilyhave the solution.
And I would say that supply chains are
in a very specific place.
They're really now at the center
of the corporate strategyand decision making.
There's a lot of excitement going on
(42:07):
and I see a lot of experimentationhappening, you know,
particularly with new products
and services, how do wedifferentiate from our competitors?
How do we use supplychain design to do that?
I do oversee that this experimentation is
not necessarily always supported
by this data-driven approach and
that the level of decisionmaking maturity is often limited.
(42:28):
And so I'm expecting that
in next few years we'llkind of see more clearly
what are some of the winningstrategies among those range
of different things thatpeople are trying out.
And basically identifiedkind of the winners
and losers of this process.
And there we will thinkthat having an intentional,
well-structured andanalytics driven approach
is gonna be key to being
in the winning side of the equation.
(42:51):
- Excellent. So how dopeople get the ball rolling?
If I'm an operator
in a company, how do Iget the ball rolling?
- Yeah, so obviously wework with a lot of companies
that are just getting started and
so what we typically wanna tell them to do
or how we coach them throughthat process is pick something
that's small but important to the business
so you can kind of work on something
(43:11):
that clearly is gonnademonstrate value to the business
and kind of show some of that value
as far as the effort that we do.
We certainly prefer somethingthat has some repeatability,
like we wanna get answers,you know, more frequently
and we wanna then build on top of
that kind of a repeatable process.
So let's get in place something
(43:32):
that not only answersquestions, but answers it
in a repeatable way so customers can begin
to see kind of how that, thevalue of that repeatability
and how it can speed a lot of the speed
to answer a lot of the questions.
Certainly there's manyorganizations out there, you know,
CTL, Coupa, other companiesthat can provide some support
to those companies to helpthem guide them along the way.
(43:54):
We certainly haveexperience with customers
that struggle maybe a littlebit more than they should.
An external kind of partnercan help kind us speed
that, you know, time into delivery.
I think overall being moreefficient for everyone.
And then of course,
leverage the existing solutions out there.
There's plenty of solutions out there
that already can kindof solve these problems.
(44:14):
Take advantage of what's outthere and make sure you fit it
to what you need, butyou know, find something
that's the right fit for youruse case and and get going.
- So Mike, Milena, thank youvery much for being here today.
- Thank you.
- Thank you. My pleasure.
Glad to participate.
- We really appreciate yourtime and super valuable to us
and hopefully it'll bevaluable to everyone.
(44:36):
So yeah, thank you.
- Thank you. Yeah, awesome.
- Thank you so much.
(upbeat music)
- All right everyone,thank you for listening.
I hope you've enjoyed this edition
of, "MIT Supply Chain Frontiers."
My name is Arthur Grau,Communications Officer
for the center, and I invite you
to visit us anytime atctl.mit.edu or search
(44:56):
for, "MIT Supply Chain Frontiers,"
on your favorite listening platform.
Until next time.
(upbeat music)
Popular Podcasts
Amy Robach & T.J. Holmes present: Aubrey O’Day, Covering the Diddy Trial
Introducing… Aubrey O’Day Diddy’s former protege, television personality, platinum selling music artist, Danity Kane alum Aubrey O’Day joins veteran journalists Amy Robach and TJ Holmes to provide a unique perspective on the trial that has captivated the attention of the nation. Join them throughout the trial as they discuss, debate, and dissect every detail, every aspect of the proceedings. Aubrey will offer her opinions and expertise, as only she is qualified to do given her first-hand knowledge. From her days on Making the Band, as she emerged as the breakout star, the truth of the situation would be the opposite of the glitz and glamour. Listen throughout every minute of the trial, for this exclusive coverage. Amy Robach and TJ Holmes present Aubrey O’Day, Covering the Diddy Trial, an iHeartRadio podcast.
Betrayal: Season 4
Karoline Borega married a man of honor – a respected Colorado Springs Police officer. She knew there would be sacrifices to accommodate her husband’s career. But she had no idea that he was using his badge to fool everyone. This season, we expose a man who swore two sacred oaths—one to his badge, one to his bride—and broke them both. We follow Karoline as she questions everything she thought she knew about her partner of over 20 years. And make sure to check out Seasons 1-3 of Betrayal, along with Betrayal Weekly Season 1.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.