May 27, 2026 • 23 mins
76% of leaders say data-driven decision making is the goal, but most people still don’t trust the data they’re looking at. That contradiction is not just frustrating, it’s expensive. We talk with Susan Walsh, founder of The Classification Guru, about what actually breaks procurement data and supplier master data over time, and why “just add AI” won’t fix a messy foundation.
We get practical about data quality in supply chain management: why cleaning and standardizing data gets treated like a side task, how the long tail of spend hides the biggest problems, and why tariffs and supply chain relocation make accurate, up-to-date data even more urgent for scenario modeling, forecasting, and real-time visibility. Susan also shares how to think about buying technology the smart way: start with your end goal, avoid paying for add-ons you don’t need, and choose tools that fit your specific use case instead of copying competitors.
Then we dig into AI, gen AI, and agentic AI. Since every model learns from training data, bad inputs can create confident-looking misinformation and spread it across your systems. We also cover data governance basics that matter globally, like consistent units of measure, date formats, naming standards, and the people-side change management that keeps data clean after the project ends.
If this conversation helps, subscribe, share it with someone wrestling with spend analytics or master data management, and leave a review so more supply chain teams can find it.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_01 (00:07):
Hey, glad to catch you no matter what you're in the
middle of doing today.
Just uh thrilled that youdecided to take some time out
and hang with us.
I'm Melanie with ISM, and yep,another episode of Supply Chain
Unfiltered.
Today we are focusing ourefforts on data.
And I would say a few years ago,you know, a lot of companies
(00:29):
were looking for data analysts,people that specialize in data.
And now it seems that almost nomatter what your role is within
a company, you're expected to beable to manage data, to analyze
data, because it's at the cruxof everything.
I want to share something withyou I ran across before I
introduce our guest.
It's a 2025 study by Preciselyand Drexel University's Labo
(00:55):
College of Business.
And in the study, they foundthat 76% of the participants
said that data-driven decisionmaking is their number one goal.
But here's the here's the otherside of that.
Of those participants statingthat, 67% of them said that they
don't trust data.
Yeah, that's a problem, right?
(01:18):
So uh we're gonna kind of headinto everything data, and who
better to chat about it than uhguest Susan Walsh.
She's the founder of theclassification guru and a
longtime friend of ISM.
So good to have you here, Susan.
So great to see you.
How are you?
I'm I'm great, I'm great, youknow.
(01:38):
How could I not be great withyou on the show?
Um true.
SPEAKER_02 (01:42):
And we're talking on Friday afternoon, so that's
always, you know, we're cool, wecan sniff the weekend.
SPEAKER_01 (01:48):
Yeah, yeah, it's just the teaser, the teaser is
here.
So um I want to kind of setthings up a little bit, uh, you
know, before I let you kind ofdive in.
But when you think about it, ifwe all look at how we're
operating today compared to fiveyears ago, and I don't need to
spell out what happened fiveyears ago.
We all know and we lived it.
(02:09):
Um but I'm wondering with allthe things that have changed in
the last five years, uh it stillseems like thing, you know, I
I'd say operationally speaking,and data has a lot to do with
it, things are kind of out ofcontrol.
I don't know if it's an again orjust kind of continuing on.
(02:30):
But how should businesses berevisiting and focusing on their
data today and what, ifanything, has changed since five
years ago?
SPEAKER_02 (02:41):
Yeah, a little bit.
Actually, coincidentally, it'sour eight-year anniversary for
classification guru this month.
So I've just uh scheduled up andbuilt a post around what is
different now from eight yearsago.
So even rolling back a littlebit further, procurement analyst
wasn't a job back then, itdidn't exist, like a lot of the
(03:05):
spend analytics was alloutsourced, and now that's all
in-house.
Um, and so when they took a lotof those skills in-house,
there's very qualified um andcapable data analysts who can do
amazing coding, they can dofantastic scripts and AI and
(03:28):
dashboards and analysis, butnobody really gets trained on
the cleaning data part.
And you know, I speak to people,and this is this is a
conversation that's been goingon since we were paper-based,
like pre-computers.
Like data quality standards havealways been an issue.
Um, unfortunately, somethingthat hasn't changed in all that
(03:50):
time is the focus and priorityof it.
I think it's kind of seen assomething less important, almost
menial as a task, but actuallygetting someone who is skilled
in it, you know, can reallyimprove and enhance your data
quality really quickly to helpyou get to where you want to go.
But all guess what?
(04:12):
Like we're going around incircles, you know.
There was new tech like fiveyears ago, and it needed clean
data to work.
All these like spend analyticstools out there, all that work,
it needed clean data to work on.
Now we've got AI, gen AI,agentic AI.
Guess what?
It needs clean data to actuallywork properly.
Nothing has changed.
(04:33):
And and I really hope, and I amseeing people talk more about
data quality now.
So I think maybe this might bethe turning point with with the
use of AI now.
unknown (04:43):
Yeah.
SPEAKER_02 (04:43):
But we'll see.
SPEAKER_01 (04:44):
Yeah, and we'll get to AI a little bit in a little
bit, but I wanted uh um get yourthoughts on tariffs.
And and um, you know, a lot ofwe're relocating some of our
supply chains for variousreasons, but um so I'm wondering
with with those two things goingon, is it a good time to upgrade
(05:06):
our data systems, or is itbetter to hold off until some of
this geopolitical dust settles?
SPEAKER_02 (05:15):
This is the best time.
I have literally had projectsput on hold because of the
tariffs, and actually that's theworst thing that could have
happened because you needaccurate, up-to-date information
and real-time visibility to beable to do like modeling on
(05:35):
what-if scenarios, um,forecasting for the future,
looking at historicalinformation, um, making
decisions based on thatinformation.
How can you compare your data ifit's not accurate or you don't
even have that information?
So, like really you are you'reon the back foot before you've
(05:56):
even started with the tariffs ifyou don't have clean and
up-to-date data.
It's going to be a and it mightnot be a problem that people
realise until it's too late aswell.
You know, it's one of thosethings you really need to be on
top of.
And again, it's nobody, apartfrom the people who really are
in the weeds and do it dealingwith it, and the the more senior
(06:18):
levels of management don't seemto appreciate the the importance
of investing in your dataquality, and it really is an
investment, it's not a cost.
SPEAKER_01 (06:27):
And I have a question.
I'm wondering, you know, withthat kind of mentality or
prioritizing those kind ofdecisions, is it that people
have an aversion to doing thedata cleaning because they'd
rather just, oh, it'll be fineand we're just going to move
forward and um focus on you knowintegrating um, you know, the
(06:53):
innovation that has all thebells and whistles, but they
don't take care of thefundamentals first.
Is that is that what goes on?
SPEAKER_02 (07:00):
Well, I I think there's definitely uh uh thought
around that it's somethingthat's really easy to do.
And the reality is we've helpedclients shave months off of
projects.
Um, you know, something that wastaken one of our university
clients a year to do, we coulddo in like three months for
(07:22):
them.
Oh like what is the value ofthat to the business?
And and that isn't somethingthat they could have just done
themselves, it was taking them ayear because they have all the
other things to do too.
We do this day in, day out, andall the other suppliers out
there that do it, you know, thatthat's really important.
And then the other thing is Ithink that most people think
(07:44):
their data is so bad and it'ssuch an overwhelming task that
they're like, you know, la la lala la la la.
No plus, you know, let's justpretend it's not happening,
let's just keep moving forward.
And and I can tell you now thatnine out of ten clients that we
work with, something has gonewrong massively.
There's trouble, and they needto fix it yesterday.
(08:05):
And that's the only time thatthe money actually becomes
available to clean the data,which is such a shame because it
should really be about beingproactive, not reactive.
SPEAKER_01 (08:14):
So that must be a lot of, I would imagine, a lot
of stress on you because yourclients, uh, I would imagine
many of them come to you in firedrill mode.
SPEAKER_02 (08:25):
Yeah.
Yeah.
Um and you have to like, I mean,with me, what you see is what
you get.
I'll be honest.
If it's gonna take, I mean, ourlongest project normally takes
about four months, but if it'sgonna take four months and not
four weeks, I'm gonna let youknow that.
Um, I always try to overoverpromise and underdeliver.
(08:48):
No, yeah, that's the one.
SPEAKER_01 (08:50):
Um wait, and you you overpromise and under deliver or
the opposite?
No, over underpromise andover-deliver.
That's the one you mean.
Yeah, yeah.
Let's go with that.
I know.
SPEAKER_02 (09:01):
But yeah, so basically, I will always add in
buffer time to each projectbecause I know things can go
wrong.
So nine out of ten times we candeliver before the the actual
estimated finished date ratherthan after it.
So try to manage theexpectations of the clients and
be honest with them about howmuch work is actually involved.
(09:24):
And sometimes it's about doinglike a little proof of concept
and showing them like thedifference between what their
data is and what it should looklike.
And sometimes that's the onlytime they kind of realize that
they've got an issue.
SPEAKER_01 (09:36):
Yeah.
SPEAKER_02 (09:37):
Oh, yeah.
Aha moments are good.
SPEAKER_01 (09:40):
Um, yeah.
SPEAKER_02 (09:40):
So also, sorry, no, um, just one more thing.
So a lot of people will kind ofmaybe cheat and look at their
top 80% of spend or suppliers oryou know, their top top
companies or top products orwhatever, and and if they're
looking okay, then they think,okay, that's fine.
But actually, in that bottomlittle bit, the tail, there is
(10:01):
so much trouble hiding in there,you really need to check it and
be thorough.
SPEAKER_01 (10:06):
I and so I wanna I wanna kind of head back into AI
for a minute and and techintegration.
And with all these things goingon, um I I imagine it's hard for
decision makers to kind ofdecipher where they should put
the spend on this type of uh onthe innovation.
(10:28):
And because there's all thesenew products, lots of new
capabilities, they all roll outso quickly, and you don't want
to invest in something thatmight end up being antiquated
sooner than later.
Um so in that vein, is there oneconsideration that you can offer
up um that businesses shouldtake a look at before they shell
(10:52):
out those dollars?
SPEAKER_02 (10:54):
Absolutely, and that is is it fit for what the
purpose that you need it?
So I hear a lot about uh withwith clients we speak to who
have committed to a three, four,or five-year deal with a with a
company, and then they'reactually paying for like half of
these add-ons that they don'teven need.
(11:15):
So I think it's really importantbefore you even set out looking
at suppliers, look at what yourend goal is.
What do you need to achievethat?
And and try not to fall into thetrap of being sold, you know,
this magical tool that's gonnado everything for you.
It doesn't exist really.
And what is right for yourorganization might be very
(11:36):
different to the next one, tothe next one, like don't just
follow what your competitors aredoing.
That's not gonna work either.
SPEAKER_01 (11:41):
Good point, good point.
Um, do you think part of thattoo is where you end up um
committing yourself to more,I'll say more tools than you
need.
Could that be uh in a way wherea company's looking at their
longer-term goals and they thinkthat, hey, you know what, if we
invest in this now, we're kindof ahead of the game because
(12:03):
we'll be ready once we reachthis goal and this one and this
one.
So they think they're beingsmart, but what if, you know, as
we know, things happen, theydon't always go according to
plan.
SPEAKER_02 (12:17):
Most companies I speak to are not thinking of the
end goal, they're thinking ofthe we need to fix something
right now.
And so the first conversation Ihave is let's not just think
about the right now, let's thinkabout longer term.
What can we build in now thatcan be useful for you in the
next six, twelve, eighteenmonths?
And I think that is where someof the problem lies too.
(12:39):
When you're having thesediscussions uh with providers,
they're telling you all thesesolutions and problems they can
solve.
But do you have any of thoseproblems?
Is that actually what you need?
Like, what do you need to besuccessful in the next three to
five years?
Will this tool help you getthere?
And I don't think mostorganizations are thinking like
(13:01):
that.
SPEAKER_01 (13:02):
And regarding AI and all the iterations of AI, yeah.
Um do you think businessesshould still be exercising
caution?
SPEAKER_02 (13:17):
Oh, absolutely.
Um, more important than ever,every single piece of AI that we
use, whether it's in the home orat work, is built on training
data.
So that training data has comefrom somewhere else.
So it has to be clean, it has tobe accurate, otherwise what the
AI is learning is not right andcould actually be contaminating
(13:40):
and populating your data withmisinformation.
And if you've got your ownin-house model, that's great
because you know exactly wherethe data's come from.
You know, you'll know if thequality is good or bad and and
and be able to judge it.
But what if that information hasbeen sourced from the internet?
Where has that come from?
How can you verify that?
(14:01):
You know, you really need to beable to sense check it and make
sure that you know it kind oflooks like what you expect it to
in the areas that you arefamiliar with.
So get an expert in a specificarea to have a look at the data
and say, does this look right toyou?
SPEAKER_01 (14:19):
Intuitiveness is a good thing too, right?
Yeah.
SPEAKER_02 (14:22):
And just experience, like we we say it all the time.
We we will spend a day on aclient's data, and by the end of
that day, we'll start to knowwhere things don't look right,
numbers might be elevated orlower than they should be, or
the description doesn't reallytie in with everything else
around it.
So, you know, it's it's a a goodskill to have and to use, but
(14:44):
definitely.
SPEAKER_01 (14:46):
So as we uh work together and we become more
integrated across the supplychains, and we get that greater
visibility and transparency thatwe all want, is it at the same
time making data management morecomplex?
(15:06):
Is there more risk in them?
SPEAKER_02 (15:09):
Um and and something that that I talk about is my
data quote.
So making sure your data isconsistent, organized, accurate,
and then it's trustworthy.
And that C, that consistency islike really important.
So if you're working with aglobal data set, is is the
language all all uniform?
Are we all using the same unitsof measure, the same date
(15:31):
formats?
Um, I think sometimes we make itmore complicated than it needs
to be by not defining standardsand and again thinking about
what do we need from not justour department or business
office or but the companyglobally, each country, you
know.
Um, and so that's where theproblems start to arise.
(15:56):
And because you there's toolsout there that can standardize
all that data, but if you stillhave people inputting and it's
wrong, that's not that's alwaysgoing to be a problem.
SPEAKER_01 (16:06):
So let's say we get to this place where we've
simplified data management andum we're looking to through that
to amplify our performance, thatwould be ideal.
But in between the time where ittakes to get to that state, uh
is that comp is that goalcompromised because not
(16:28):
everybody is operating from thatsame foundation that you just
mentioned?
Where um and is that going totake more time?
Like how how long is it gonnatake before we're all in that
same same place, I'll say, ofdata nirvana?
SPEAKER_02 (16:48):
Well, that's a whole change management piece.
Yeah, that's and that's a peopleproblem.
So I do say data problems arepeople problems on the whole,
and that is because you know, inespecially with procurement data
and supply chain data, you canhave more than one right answer.
And again, it's about beingconsistent, having the same
(17:09):
thing.
And if you have two people whoare arguing because they're both
right, and one is always goingto input data as this A, and one
is always going to input it asB, then you're gonna have
problems.
So it's about communication,education, um giving them
responsibility as well for thedata that they're managing and
(17:30):
and telling them how it impactsthe rest of the business when
it's not right.
Yeah, it absolutely can likehold things up, and and it
depends on the type of data, itdepends on within the
organization, like you know, howimportant is that piece of part
of the data to the rest of theorganization?
(17:51):
You know, if it's I've tried tothink of something that wouldn't
be important, but uh oh, it'sall important.
I don't know, someone's shoesize, right?
Okay, so you need a shoe size,someone's shoe size for I don't
know, uniform or something thatthat's not gonna be as important
as uh categorizing your suppliercorrectly, giving them the right
GL code, um, giving them acorrect bank account and head
(18:16):
office number, tax ID, thosekinds of things.
So, you know, it's gonna dependas well.
SPEAKER_01 (18:22):
Okay.
So I know you know everything,you are the queen of data.
So I want to know, and I wantyou to share with us is there
something going on that weshould get excited about?
Um, what's coming?
You know, maybe there's aspecific program or methodology
that's gonna make our liveseasier, uh, something you're
(18:43):
working on that you want toshare.
SPEAKER_02 (18:45):
Well, I mean, I'll start with me because why not?
But um, I have built a toolcalled Samification, and at the
moment it normalizes suppliernames.
You literally sign into thewebsite, drop your file in, and
within anything from 10 minutesto two hours, you get a file
back depending on the volume.
Nice, and that's based on a400,000 supplier master list
(19:09):
that we have, so we matcheverything.
Anything that doesn't match thengoes through a set of rules and
our own chat GPT model.
So it's getting really good, butvery exciting.
Um, in the next few weeks, we'regoing to start launching an
add-on, which is two-levelgeneric classification per
supplier, just as uh a bit ofinformation.
(19:30):
But I know the thing that theprocurement are asking for a lot
is the parent companyinformation and not from a
finance point of view where theyneed to know like from a risk
perspective, but more from a wejust need to know who the parent
company is so we know who to goto negotiate with at the top
level.
So we're going to try to workover the summer and try and
(19:50):
build that in as an add-on aswell.
But I think us aside, I thinkagencai is really gonna take
over and dominate in the next 12months.
It's already all the events thatI go to, people are already
talking about it, but I've seensome amazing use cases already.
And I mean, literally the onlything that's stopping you is
(20:13):
your imagination with thesethings, you know, it can it's
gonna really help streamline alot of processes and take away
some of that manual effort ofjust you know pulling it from
this spreadsheet and putting itinto this spreadsheet, and then
it might make a decision, youknow, it can make decisions and
things if you put the rightinformation in.
So again, it's all gonna comedown to accurate data, but it
(20:38):
could truly transform all ourlives.
I mean, I'm thinking about how Ican use it more in my business.
SPEAKER_01 (20:43):
Hmm.
That's pretty exciting.
Uh it's and you know, it's it'ssuper important to stay
connected to the people that canshare these updates, like
yourself.
Uh I I I guess for uh for anyoneum tuning in that um wants to
(21:04):
connect with you, if you haven'tsigned up for Susan's
newsletter, you still have thatgoing on, right?
SPEAKER_02 (21:10):
I do.
Okay.
SPEAKER_01 (21:12):
It's always really chunky as well.
There's always a lot going on.
So if you'd like to uh getconnected with her newsletter or
want to just you know have yourown conversation with Susan, how
can they do that?
SPEAKER_02 (21:24):
LinkedIn is the best place for me, Susan Walsh, the
classification guru.
Um, if we connect or if youfollow me, you'll get an invite
to the newsletter so you don'teven have to do anything.
Um I you can go totheclassificationguru.com and
have a look around,samification.com.
Um, and yeah, we answer anyquestions.
(21:44):
You my email address is on theretoo, Susan at
theclassificationguru.com.
So yeah.
SPEAKER_01 (21:50):
So, so glad that you were able to carve out some time
for us.
Uh really great to catch up.
And um that's great.
And I'm so jealous of youroffer.
You put mine to shame.
Oh well, um yeah, but you know,at least at least I'm I'm
venturing to guess where youare.
(22:10):
It's not, it doesn't get to 115degrees out.
I don't know what that convertsto in Celsius, but let me tell
you, it's hot.
SPEAKER_02 (22:18):
Um, it's well, you say that it's actually we're
having our hottest day of theyear today.
Really?
But you guys, you guys havenailed the aircon.
We don't really have airconhere, so like tonight it's going
to be sweaty for me.
SPEAKER_01 (22:33):
Really?
Okay.
Yeah.
Well, go get yourself a coolbeverage.
Thank you.
Thank you for hanging out.
Really appreciate it.
Um, to catch any of our shows,feel free to check out the
library of content,ismworld.org.
That's where you'll find it,along with a lot of great uh
resources to keep your businessin the know and on track to
(22:54):
achieve the goals that uh you'relooking forward to doing.
Thanks for tuning in.
Always a pleasure.
I'm Melanie Stern for ISM.
Hey, glad to catch you no matter what you're in the
middle of doing today.
Just uh thrilled that youdecided to take some time out
and hang with us.
I'm Melanie with ISM, and yep,another episode of Supply Chain
Unfiltered.
Today we are focusing ourefforts on data.
And I would say a few years ago,you know, a lot of companies
(00:29):
were looking for data analysts,people that specialize in data.
And now it seems that almost nomatter what your role is within
a company, you're expected to beable to manage data, to analyze
data, because it's at the cruxof everything.
I want to share something withyou I ran across before I
introduce our guest.
It's a 2025 study by Preciselyand Drexel University's Labo
(00:55):
College of Business.
And in the study, they foundthat 76% of the participants
said that data-driven decisionmaking is their number one goal.
But here's the here's the otherside of that.
Of those participants statingthat, 67% of them said that they
don't trust data.
Yeah, that's a problem, right?
(01:18):
So uh we're gonna kind of headinto everything data, and who
better to chat about it than uhguest Susan Walsh.
She's the founder of theclassification guru and a
longtime friend of ISM.
So good to have you here, Susan.
So great to see you.
How are you?
I'm I'm great, I'm great, youknow.
(01:38):
How could I not be great withyou on the show?
Um true.
SPEAKER_02 (01:42):
And we're talking on Friday afternoon, so that's
always, you know, we're cool, wecan sniff the weekend.
SPEAKER_01 (01:48):
Yeah, yeah, it's just the teaser, the teaser is
here.
So um I want to kind of setthings up a little bit, uh, you
know, before I let you kind ofdive in.
But when you think about it, ifwe all look at how we're
operating today compared to fiveyears ago, and I don't need to
spell out what happened fiveyears ago.
We all know and we lived it.
(02:09):
Um but I'm wondering with allthe things that have changed in
the last five years, uh it stillseems like thing, you know, I
I'd say operationally speaking,and data has a lot to do with
it, things are kind of out ofcontrol.
I don't know if it's an again orjust kind of continuing on.
(02:30):
But how should businesses berevisiting and focusing on their
data today and what, ifanything, has changed since five
years ago?
SPEAKER_02 (02:41):
Yeah, a little bit.
Actually, coincidentally, it'sour eight-year anniversary for
classification guru this month.
So I've just uh scheduled up andbuilt a post around what is
different now from eight yearsago.
So even rolling back a littlebit further, procurement analyst
wasn't a job back then, itdidn't exist, like a lot of the
(03:05):
spend analytics was alloutsourced, and now that's all
in-house.
Um, and so when they took a lotof those skills in-house,
there's very qualified um andcapable data analysts who can do
amazing coding, they can dofantastic scripts and AI and
(03:28):
dashboards and analysis, butnobody really gets trained on
the cleaning data part.
And you know, I speak to people,and this is this is a
conversation that's been goingon since we were paper-based,
like pre-computers.
Like data quality standards havealways been an issue.
Um, unfortunately, somethingthat hasn't changed in all that
(03:50):
time is the focus and priorityof it.
I think it's kind of seen assomething less important, almost
menial as a task, but actuallygetting someone who is skilled
in it, you know, can reallyimprove and enhance your data
quality really quickly to helpyou get to where you want to go.
But all guess what?
(04:12):
Like we're going around incircles, you know.
There was new tech like fiveyears ago, and it needed clean
data to work.
All these like spend analyticstools out there, all that work,
it needed clean data to work on.
Now we've got AI, gen AI,agentic AI.
Guess what?
It needs clean data to actuallywork properly.
Nothing has changed.
(04:33):
And and I really hope, and I amseeing people talk more about
data quality now.
So I think maybe this might bethe turning point with with the
use of AI now.
unknown (04:43):
Yeah.
SPEAKER_02 (04:43):
But we'll see.
SPEAKER_01 (04:44):
Yeah, and we'll get to AI a little bit in a little
bit, but I wanted uh um get yourthoughts on tariffs.
And and um, you know, a lot ofwe're relocating some of our
supply chains for variousreasons, but um so I'm wondering
with with those two things goingon, is it a good time to upgrade
(05:06):
our data systems, or is itbetter to hold off until some of
this geopolitical dust settles?
SPEAKER_02 (05:15):
This is the best time.
I have literally had projectsput on hold because of the
tariffs, and actually that's theworst thing that could have
happened because you needaccurate, up-to-date information
and real-time visibility to beable to do like modeling on
(05:35):
what-if scenarios, um,forecasting for the future,
looking at historicalinformation, um, making
decisions based on thatinformation.
How can you compare your data ifit's not accurate or you don't
even have that information?
So, like really you are you'reon the back foot before you've
(05:56):
even started with the tariffs ifyou don't have clean and
up-to-date data.
It's going to be a and it mightnot be a problem that people
realise until it's too late aswell.
You know, it's one of thosethings you really need to be on
top of.
And again, it's nobody, apartfrom the people who really are
in the weeds and do it dealingwith it, and the the more senior
(06:18):
levels of management don't seemto appreciate the the importance
of investing in your dataquality, and it really is an
investment, it's not a cost.
SPEAKER_01 (06:27):
And I have a question.
I'm wondering, you know, withthat kind of mentality or
prioritizing those kind ofdecisions, is it that people
have an aversion to doing thedata cleaning because they'd
rather just, oh, it'll be fineand we're just going to move
forward and um focus on you knowintegrating um, you know, the
(06:53):
innovation that has all thebells and whistles, but they
don't take care of thefundamentals first.
Is that is that what goes on?
SPEAKER_02 (07:00):
Well, I I think there's definitely uh uh thought
around that it's somethingthat's really easy to do.
And the reality is we've helpedclients shave months off of
projects.
Um, you know, something that wastaken one of our university
clients a year to do, we coulddo in like three months for
(07:22):
them.
Oh like what is the value ofthat to the business?
And and that isn't somethingthat they could have just done
themselves, it was taking them ayear because they have all the
other things to do too.
We do this day in, day out, andall the other suppliers out
there that do it, you know, thatthat's really important.
And then the other thing is Ithink that most people think
(07:44):
their data is so bad and it'ssuch an overwhelming task that
they're like, you know, la la lala la la la.
No plus, you know, let's justpretend it's not happening,
let's just keep moving forward.
And and I can tell you now thatnine out of ten clients that we
work with, something has gonewrong massively.
There's trouble, and they needto fix it yesterday.
(08:05):
And that's the only time thatthe money actually becomes
available to clean the data,which is such a shame because it
should really be about beingproactive, not reactive.
SPEAKER_01 (08:14):
So that must be a lot of, I would imagine, a lot
of stress on you because yourclients, uh, I would imagine
many of them come to you in firedrill mode.
SPEAKER_02 (08:25):
Yeah.
Yeah.
Um and you have to like, I mean,with me, what you see is what
you get.
I'll be honest.
If it's gonna take, I mean, ourlongest project normally takes
about four months, but if it'sgonna take four months and not
four weeks, I'm gonna let youknow that.
Um, I always try to overoverpromise and underdeliver.
(08:48):
No, yeah, that's the one.
SPEAKER_01 (08:50):
Um wait, and you you overpromise and under deliver or
the opposite?
No, over underpromise andover-deliver.
That's the one you mean.
Yeah, yeah.
Let's go with that.
I know.
SPEAKER_02 (09:01):
But yeah, so basically, I will always add in
buffer time to each projectbecause I know things can go
wrong.
So nine out of ten times we candeliver before the the actual
estimated finished date ratherthan after it.
So try to manage theexpectations of the clients and
be honest with them about howmuch work is actually involved.
(09:24):
And sometimes it's about doinglike a little proof of concept
and showing them like thedifference between what their
data is and what it should looklike.
And sometimes that's the onlytime they kind of realize that
they've got an issue.
SPEAKER_01 (09:36):
Yeah.
SPEAKER_02 (09:37):
Oh, yeah.
Aha moments are good.
SPEAKER_01 (09:40):
Um, yeah.
SPEAKER_02 (09:40):
So also, sorry, no, um, just one more thing.
So a lot of people will kind ofmaybe cheat and look at their
top 80% of spend or suppliers oryou know, their top top
companies or top products orwhatever, and and if they're
looking okay, then they think,okay, that's fine.
But actually, in that bottomlittle bit, the tail, there is
(10:01):
so much trouble hiding in there,you really need to check it and
be thorough.
SPEAKER_01 (10:06):
I and so I wanna I wanna kind of head back into AI
for a minute and and techintegration.
And with all these things goingon, um I I imagine it's hard for
decision makers to kind ofdecipher where they should put
the spend on this type of uh onthe innovation.
(10:28):
And because there's all thesenew products, lots of new
capabilities, they all roll outso quickly, and you don't want
to invest in something thatmight end up being antiquated
sooner than later.
Um so in that vein, is there oneconsideration that you can offer
up um that businesses shouldtake a look at before they shell
(10:52):
out those dollars?
SPEAKER_02 (10:54):
Absolutely, and that is is it fit for what the
purpose that you need it?
So I hear a lot about uh withwith clients we speak to who
have committed to a three, four,or five-year deal with a with a
company, and then they'reactually paying for like half of
these add-ons that they don'teven need.
(11:15):
So I think it's really importantbefore you even set out looking
at suppliers, look at what yourend goal is.
What do you need to achievethat?
And and try not to fall into thetrap of being sold, you know,
this magical tool that's gonnado everything for you.
It doesn't exist really.
And what is right for yourorganization might be very
(11:36):
different to the next one, tothe next one, like don't just
follow what your competitors aredoing.
That's not gonna work either.
SPEAKER_01 (11:41):
Good point, good point.
Um, do you think part of thattoo is where you end up um
committing yourself to more,I'll say more tools than you
need.
Could that be uh in a way wherea company's looking at their
longer-term goals and they thinkthat, hey, you know what, if we
invest in this now, we're kindof ahead of the game because
(12:03):
we'll be ready once we reachthis goal and this one and this
one.
So they think they're beingsmart, but what if, you know, as
we know, things happen, theydon't always go according to
plan.
SPEAKER_02 (12:17):
Most companies I speak to are not thinking of the
end goal, they're thinking ofthe we need to fix something
right now.
And so the first conversation Ihave is let's not just think
about the right now, let's thinkabout longer term.
What can we build in now thatcan be useful for you in the
next six, twelve, eighteenmonths?
And I think that is where someof the problem lies too.
(12:39):
When you're having thesediscussions uh with providers,
they're telling you all thesesolutions and problems they can
solve.
But do you have any of thoseproblems?
Is that actually what you need?
Like, what do you need to besuccessful in the next three to
five years?
Will this tool help you getthere?
And I don't think mostorganizations are thinking like
(13:01):
that.
SPEAKER_01 (13:02):
And regarding AI and all the iterations of AI, yeah.
Um do you think businessesshould still be exercising
caution?
SPEAKER_02 (13:17):
Oh, absolutely.
Um, more important than ever,every single piece of AI that we
use, whether it's in the home orat work, is built on training
data.
So that training data has comefrom somewhere else.
So it has to be clean, it has tobe accurate, otherwise what the
AI is learning is not right andcould actually be contaminating
(13:40):
and populating your data withmisinformation.
And if you've got your ownin-house model, that's great
because you know exactly wherethe data's come from.
You know, you'll know if thequality is good or bad and and
and be able to judge it.
But what if that information hasbeen sourced from the internet?
Where has that come from?
How can you verify that?
(14:01):
You know, you really need to beable to sense check it and make
sure that you know it kind oflooks like what you expect it to
in the areas that you arefamiliar with.
So get an expert in a specificarea to have a look at the data
and say, does this look right toyou?
SPEAKER_01 (14:19):
Intuitiveness is a good thing too, right?
Yeah.
SPEAKER_02 (14:22):
And just experience, like we we say it all the time.
We we will spend a day on aclient's data, and by the end of
that day, we'll start to knowwhere things don't look right,
numbers might be elevated orlower than they should be, or
the description doesn't reallytie in with everything else
around it.
So, you know, it's it's a a goodskill to have and to use, but
(14:44):
definitely.
SPEAKER_01 (14:46):
So as we uh work together and we become more
integrated across the supplychains, and we get that greater
visibility and transparency thatwe all want, is it at the same
time making data management morecomplex?
(15:06):
Is there more risk in them?
SPEAKER_02 (15:09):
Um and and something that that I talk about is my
data quote.
So making sure your data isconsistent, organized, accurate,
and then it's trustworthy.
And that C, that consistency islike really important.
So if you're working with aglobal data set, is is the
language all all uniform?
Are we all using the same unitsof measure, the same date
(15:31):
formats?
Um, I think sometimes we make itmore complicated than it needs
to be by not defining standardsand and again thinking about
what do we need from not justour department or business
office or but the companyglobally, each country, you
know.
Um, and so that's where theproblems start to arise.
(15:56):
And because you there's toolsout there that can standardize
all that data, but if you stillhave people inputting and it's
wrong, that's not that's alwaysgoing to be a problem.
SPEAKER_01 (16:06):
So let's say we get to this place where we've
simplified data management andum we're looking to through that
to amplify our performance, thatwould be ideal.
But in between the time where ittakes to get to that state, uh
is that comp is that goalcompromised because not
(16:28):
everybody is operating from thatsame foundation that you just
mentioned?
Where um and is that going totake more time?
Like how how long is it gonnatake before we're all in that
same same place, I'll say, ofdata nirvana?
SPEAKER_02 (16:48):
Well, that's a whole change management piece.
Yeah, that's and that's a peopleproblem.
So I do say data problems arepeople problems on the whole,
and that is because you know, inespecially with procurement data
and supply chain data, you canhave more than one right answer.
And again, it's about beingconsistent, having the same
(17:09):
thing.
And if you have two people whoare arguing because they're both
right, and one is always goingto input data as this A, and one
is always going to input it asB, then you're gonna have
problems.
So it's about communication,education, um giving them
responsibility as well for thedata that they're managing and
(17:30):
and telling them how it impactsthe rest of the business when
it's not right.
Yeah, it absolutely can likehold things up, and and it
depends on the type of data, itdepends on within the
organization, like you know, howimportant is that piece of part
of the data to the rest of theorganization?
(17:51):
You know, if it's I've tried tothink of something that wouldn't
be important, but uh oh, it'sall important.
I don't know, someone's shoesize, right?
Okay, so you need a shoe size,someone's shoe size for I don't
know, uniform or something thatthat's not gonna be as important
as uh categorizing your suppliercorrectly, giving them the right
GL code, um, giving them acorrect bank account and head
(18:16):
office number, tax ID, thosekinds of things.
So, you know, it's gonna dependas well.
SPEAKER_01 (18:22):
Okay.
So I know you know everything,you are the queen of data.
So I want to know, and I wantyou to share with us is there
something going on that weshould get excited about?
Um, what's coming?
You know, maybe there's aspecific program or methodology
that's gonna make our liveseasier, uh, something you're
(18:43):
working on that you want toshare.
SPEAKER_02 (18:45):
Well, I mean, I'll start with me because why not?
But um, I have built a toolcalled Samification, and at the
moment it normalizes suppliernames.
You literally sign into thewebsite, drop your file in, and
within anything from 10 minutesto two hours, you get a file
back depending on the volume.
Nice, and that's based on a400,000 supplier master list
(19:09):
that we have, so we matcheverything.
Anything that doesn't match thengoes through a set of rules and
our own chat GPT model.
So it's getting really good, butvery exciting.
Um, in the next few weeks, we'regoing to start launching an
add-on, which is two-levelgeneric classification per
supplier, just as uh a bit ofinformation.
(19:30):
But I know the thing that theprocurement are asking for a lot
is the parent companyinformation and not from a
finance point of view where theyneed to know like from a risk
perspective, but more from a wejust need to know who the parent
company is so we know who to goto negotiate with at the top
level.
So we're going to try to workover the summer and try and
(19:50):
build that in as an add-on aswell.
But I think us aside, I thinkagencai is really gonna take
over and dominate in the next 12months.
It's already all the events thatI go to, people are already
talking about it, but I've seensome amazing use cases already.
And I mean, literally the onlything that's stopping you is
(20:13):
your imagination with thesethings, you know, it can it's
gonna really help streamline alot of processes and take away
some of that manual effort ofjust you know pulling it from
this spreadsheet and putting itinto this spreadsheet, and then
it might make a decision, youknow, it can make decisions and
things if you put the rightinformation in.
So again, it's all gonna comedown to accurate data, but it
(20:38):
could truly transform all ourlives.
I mean, I'm thinking about how Ican use it more in my business.
SPEAKER_01 (20:43):
Hmm.
That's pretty exciting.
Uh it's and you know, it's it'ssuper important to stay
connected to the people that canshare these updates, like
yourself.
Uh I I I guess for uh for anyoneum tuning in that um wants to
(21:04):
connect with you, if you haven'tsigned up for Susan's
newsletter, you still have thatgoing on, right?
SPEAKER_02 (21:10):
I do.
Okay.
SPEAKER_01 (21:12):
It's always really chunky as well.
There's always a lot going on.
So if you'd like to uh getconnected with her newsletter or
want to just you know have yourown conversation with Susan, how
can they do that?
SPEAKER_02 (21:24):
LinkedIn is the best place for me, Susan Walsh, the
classification guru.
Um, if we connect or if youfollow me, you'll get an invite
to the newsletter so you don'teven have to do anything.
Um I you can go totheclassificationguru.com and
have a look around,samification.com.
Um, and yeah, we answer anyquestions.
(21:44):
You my email address is on theretoo, Susan at
theclassificationguru.com.
So yeah.
SPEAKER_01 (21:50):
So, so glad that you were able to carve out some time
for us.
Uh really great to catch up.
And um that's great.
And I'm so jealous of youroffer.
You put mine to shame.
Oh well, um yeah, but you know,at least at least I'm I'm
venturing to guess where youare.
(22:10):
It's not, it doesn't get to 115degrees out.
I don't know what that convertsto in Celsius, but let me tell
you, it's hot.
SPEAKER_02 (22:18):
Um, it's well, you say that it's actually we're
having our hottest day of theyear today.
Really?
But you guys, you guys havenailed the aircon.
We don't really have airconhere, so like tonight it's going
to be sweaty for me.
SPEAKER_01 (22:33):
Really?
Okay.
Yeah.
Well, go get yourself a coolbeverage.
Thank you.
Thank you for hanging out.
Really appreciate it.
Um, to catch any of our shows,feel free to check out the
library of content,ismworld.org.
That's where you'll find it,along with a lot of great uh
resources to keep your businessin the know and on track to
(22:54):
achieve the goals that uh you'relooking forward to doing.
Thanks for tuning in.
Always a pleasure.
I'm Melanie Stern for ISM.
Popular Podcasts
Hey Jonas!
Hey Jonas! The official Jonas Brothers podcast. Hosted by Kevin, Joe, and Nick Jonas. It’s the Jonas Brothers you know... musicians, actors, and well, yes, brothers. Now, they’re sharing another side of themselves in the playful, intimate, and irreverent way only they can. Spend time with the Jonas Brothers here and stay a little bit longer for deep conversations like never before.
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.
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.