December 23, 2024 • 34 mins
Join us for an insightful chat with Siqi Chen, CEO of Runway and a visionary in AI and finance, on this episode of Count Me In! Host Adam Larson sits down with Siqi to explore the incredible potential of AI in streamlining our work, making complex tasks more intuitive. Discover the fascinating world of prompt engineering and how Runway, Siqi's brainchild, is redefining tools for strategic thinking.
Siqi shares his unique journey from working on Mars rovers to revolutionizing financial planning. Dive into a discussion on the democratization of data, the evolution of finance from a backward-looking to a forward-thinking strategy, and the shortcomings of traditional tools like spreadsheets. This episode is full of compelling insights on making financial models accessible and integrating context seamlessly into complex projects.
Sponsor:
Today's episode is brought to you by U.S. Bank. U.S. Bank is a trusted financial partner for our clients, businesses and communities. We believe in doing the right thing and putting people first. It’s an honor to be recognized as one of the World’s Most Ethical Companies® by the Ethisphere Institute for the tenth consecutive year. From commercial credit cards and program management tools to innovative payment technologies and transportation offerings, U.S. Bank Corporate Payment Systems has the right solution to help your organization reduce payment costs, enhance control and streamline your entire payment processing function. We’ll partner with you to uncover your challenges and provide smart, clear and honest guidance to help you meet the financial goals for your business. Visit usbank.com/corporatepayments to learn more.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Adam Larson (00:21):
Welcome to Count Me In. In this episode, we are
joined by the innovative SiqiChen, CEO of Runway, and a 4
time CEO and 3 time founder.With a rich background in math,
technology, and productmanagement, Zeki shares
groundbreaking insight intoleveraging AI for enhanced work
efficiency and intuitivethinking tools. We discussed the
nuances of prompt engineering,Runway's revolutionary approach
(00:43):
to AI, and the importance ofdesign and financial planning.
Zeki contrast traditional toolslike Excel with cutting edge
solutions, aiming to democratizefinancial data access and
improve collaboration acrossorganizations.
Tune in for an engagingconversation that promises to
reshape how you think about AI,financial modeling, and the
future of work. Well, Siqi, I'mreally excited to have you on
(01:11):
the podcast, and we're gonna betalking a lot about, you know,
FP and a and financial planningwithin organizations. You know?
But I wanna talk a little bitabout your story because you
don't have the normal, you know,starting or running an
organization, a financialorganization, story. You're not
a typical, hey.
I started and did my, you know,did my accounting with Deloitte
(01:33):
or KPMG and, you know, kindawent up that way. You don't have
that typical story. So I wantedto talk a little bit about your
story first.
Siqi Chen (01:39):
Yeah. Happy to share. You're right. This is my first
time running a finance org andbuilding anything in B2B much
less, in finance software. Mybackground is in math.
I majored in math, in fact, andit's mostly in technology. So,
my first job, during college wasworking on the Mars rovers at
(02:02):
JPL, in machine vision. And overthe course of my career, now a 4
time CEO, 3 time founder. And Inever really felt comfortable
with finance to the extent thatI was asking our 1st head of
finance, whether I should justgo get an MBA. And he told me
no.
(02:23):
Do it. So that very muchdirectly inspired what we're,
what I'm doing now. But, youknow, as in retrospect though, a
lot of the things that I workedon, in my background is in
product and growth eventually.Right? So I ran product at a
company called Zynga.
It was, it made Farmville, whichmany people may have know and
(02:43):
now a public company. But atthat company, product management
was about modeling. It's aboutmodeling retention and how users
come in and how we monetizethem. And we did it in
spreadsheets and that reallyinformed how we think about
finance at runway. Because whatI realized is that, you know,
(03:04):
from a first principleperspective, financial models
are not actually different than,these product growth models.
They're in different units andthe terminology is different.
But at the end of day, what allof these models do is they are
simulations of something. Theylet you play forward, you know,
time travel into the future tosee what's gonna happen to the
(03:24):
business today and the decisionsthat you make today. And so I
very much, take a broader firstprinciples view of what finance
is about. I think that's beeninteresting for the product and
for our organization.
Adam Larson (03:36):
That really has. It's very it's a unique
perspective, but I think withinorganizations and with the
finance organization,especially, you kinda need to
take a different approach to,first of all, get new clients
and and make it more applicable.But how do you take, like,
things like finance with allthese different lingos and make
it understandable for, you know,finance professionals, but also
other people within the companyor even new clients? Yeah.
(03:59):
That's a great question.
One of our core beliefs is thatpools inform human behavior more
than people give credit, givethem credit for. Okay. And so
we, right now, you know, we, welive in a world where finance in
most organizations is verysiloed
Siqi Chen (04:16):
from the rest of the departments, right? It feels
like black box and thisfrustrates, not just the rest of
the company, but also thefinance leaders, right? Finance
leaders don't feel appreciatedor understood for the value that
they bring to the table, whichis substantial. And a lot of
this is lingo and it isunderstanding that it's
technical, but it's also theavailability of tooling to make
(04:39):
it accessible andunderstandable. So by way of a
simple example, let's say yourmodel lives on Google sheets or
Excel.
And you are a very forwardlooking finance person, and you
want everyone in the company tounderstand how the business
works. So you want people, youwant to share this model with
the rest of the company. Youdon't. And people think it's
because the finance leadersdon't want to. But the reason
(05:02):
why they don't want to isthey're afraid of few really
important invalid things.
The first thing is you giveeveryone access and people might
accidentally change somethingwithout, you know, that's like
legitimate. You know, you wantto have those locked in
operating plan and so am I fatfinger or something? The second
reason is there is no capabilityon Google sheets or Excel to
hide a single column. And thatcolumn is usually called
(05:24):
salaries. Right.
Because all calculation isarchitected to be done on the
front end. And so because oftools don't allow you to do it,
people don't do it. And then thefinance department and leaders
think, oh, they, they, they getthought of as, as like this,
this very protective departmentand people who are very
sensitive to who gets to seewhat, when in actuality the
(05:46):
tools and support it. And if youlook at sort of the last
generation of software fordesign and analytics, right?
Like Figma is probably a goodexample for designers before the
invention of Figma.
It was rude to look over theshoulder of a designer as they
were working. Right. But afterFigma happened because it was so
(06:10):
collaborative of so accessible,Everyone lives in Sigma, whether
you're a product person or abusiness person or engineer. And
it's okay because the productwas designed to do that. There
is no such product for financetoday.
And we think that's a hugeopportunity. And it can change
behavior. And the behavior iswhat it is because the products
don't do that. Wow.
Adam Larson (06:35):
It's like creating Figma for finance, basically
what you're saying, right?
Siqi Chen (06:40):
That is literally the tagline for our seed pitch. And
it was inspiration for why thecompany existed. In 2020, I was
running a company called sandboxVR, very complicated retail VR
hardware. I was justdisappointed CEO 6 months ago
and COVID hits and we were doingscenario planning, and very
(07:01):
quickly our revenue went from,you know, tens of 1,000,000 to
0. And we had to figure out howwe're going to survive.
We had like 400 employees, wentfrom 400 to 10 and I laid myself
off and retired executive team.And 30 minutes afterwards, I
talked to our CFO, our at CFOfrom sandbox. And I asked him,
Hey, we have really amazingtools today. We have Figma and
(07:24):
notion and air table. Surelythere was something better we
could have used to do all thescenario planning than Google
sheets.
We just didn't bother using it.And he laughed and said, no,
this is like all there is. So Italked to Andreessen team and
that's how wrong we got started.So that's actually the direct,
inspiration. And Dylan, thefounder of Figma is one of our
first investors for the samereason.
Adam Larson (07:45):
Wow.
So how is that gonna, how canthat revolutionize an
organization if you can have figa Figma for finance and
accounting team? Because a lotof times, you know, you have
these big old spreadsheets, likeyou said, Google Sheets. You
have, you know, these, ERPsystems like that. They think
they can they can do similarthings, but not exactly how
you're
Siqi Chen (08:03):
describing. Yeah. So, I think it's probably one of the
most impactful things someonecan work on for business. So one
of the, my observations is youspend 100 of 1,000 of dollars,
like implementing something likea finance platform or ERP, and
(08:25):
you pay a 100, a few $100,000.And on top of that, you pay
maybe an extra, a few 100,000hours to implement it over the
course of 6 months or 2 years,depending on the company.
And what happens is you spendall this time and you connect
all this data and you have thisforecast, but the minute the
business changes, you have tolike pay somebody again to
(08:46):
change it because you can'ttouch it. It's so delicate and
sensitive and fragile. And theother thing that happens is
let's say a department leader,like a product person or a GM or
whoever has some idea about whatthey're gonna do in the business
and they need to model it. Thelast thing they're going to
going to go to try that is goingto start with Anaplan or
(09:08):
something like that. The firstthing that you reach for is the
most convenient modeling toolthat's been ever invented, which
is Excel.
Right? And that seed of an ideathen becomes a real model. And
eventually you have aspreadsheet model in every
different department. And soyou've spent 100 and 1,000
dollars for this thing, but thereal reality, the real model of
(09:32):
how these different departmentswork lives completely outside of
that. And as a result, peoplehave misalignment and there's no
visibility.
And your great model in your FPand a platform doesn't reflect
any of that. And so if you canbuild something that is people's
first choice to think of anidea, because it's easier,
(09:54):
faster, and more intuitive, theimpact is that everyone in the
company actually gets to have areal view of the business. A
true view of the business is upto date. So when I look at the
product, you know, the, theproduct ropes part of the model
that actually represents whatthe current thinking is of the
product team, what the marketingplan is, what the sales plan is
(10:15):
instead of like sort of a, maybea 3 month delay to view entered
because I sent a spreadsheetover slack to the sales head to,
to fill it out and then paste itback in. And the the dream that
I think we're very close toachieving.
And in some of our customersalready achieved this as imagine
(10:37):
a designer or engineer or a PMbe able to say, if we work on X
instead of Y here is how it'sgoing to impact margin and
growth for our company 2 yearsfrom now.
Adam Larson (10:48):
Wow. That conversation does not happen,
which is actually wild becausewe all are working on the
business. Everything that we doimpacts the bottom line
somewhere. That is a possibleworld. And it's not possible
today because the tools simplyaren't there, but it's totally
achievable and totally billable.
Well, yeah, it seems like in intoday's world, you know, if
(11:09):
you're budgeting or or doingthese these big old strategic
planning things, it's thismonths months long processes.
You're doing spreadsheets.You're throwing things around.
Then you have meetings with thebig wigs, and then the the big
wigs talk to the peopleunderneath them, and then it's
spread out to everybody. Andit's this long process, but what
you're describing sounds like itcould be in a matter of days.
You're like, hey. This is whatwe're planning on. This is what
(11:30):
we're looking to do. And and itbreaks down the silos that are
today built up in organizations.
Siqi Chen (11:35):
That's right. I mean, we have some of our customers
integrating Jira into Runway,which, you know, is unheard of
for a finance platform. Wow. Butto your point, it's not just
about the annual quarterlybudgeting. Right?
When you're thinking about thenext few items on your roadmap,
where your next project is, youshould have a conversation and
you should have understanding oflike, what is this going to move
(11:55):
and how does it impact overallbusiness? And so one of the big
opportunities is if you thinkabout how finance has evolved as
a discipline, right? I think weall know it's evolving from this
sort of backwards lookingcontrols reporting to more of a
forward looking decision makingstrategic role. What does, what
(12:17):
does it mean to be strategic?Right?
Like it means you have to lookinto the future, but also it's
not just a forecasting thefuture accurately. It's a
understanding the causality ofwhat are the things that you can
do today to move the future.And, and that's underappreciated
and, and you need better toolsfor everyone to have that same
(12:38):
understanding here are the keydrivers that drive growth, not
just like we're good forecasting10% year over year growth or 20%
year over year growth. Anyonecan do that. Right?
But, like, it doesn't help youmake better decisions. And,
ultimately, that's whatstrategic finance is really
about, helping the entireorganization make better
decisions. So how
Adam Larson (12:56):
do you how do you talk have that conversation,
especially with people who'vebeen in within the finance
function, the accountingfunction in organizations for
years? They've always done itthe same way, and a lot of times
change comes to the finance teamlast because there's so much
that they're doing. It's hard toimplement these other things.
How do you have thatconversation to to kind of say,
hey. You can do something new?
You, I mean, yeah, our our ourmain move is to show them the
(13:19):
product. Okay.
Siqi Chen (13:21):
So if you look at our website, if you click our
videos, you know, it's verydifferent from a typical B2B
enterprise, like type ofwebsite. Yeah. Or you just have
like smiling people, likeshaking hands from stock photos.
The software, like I'm a productperson. Software is designed to
be used.
Of course. Like you have to seeit. It has to be good because
(13:43):
if, if you're not showing offthe product, you're not showing
people how it can work, thenyou're gonna buy software that
is gonna sound good on paperbecause it's really well sold,
but it actually doesn't do thething. And that's our
observation. People have beentalking about connected finance
and strategic planning for like10, 20 years.
It doesn't do the thing. And youhave to invent new product
(14:04):
abstractions and experiences inorder for it to do the thing.
And it's a much higher bar. Sothat's like one component is how
we knew it. But the othercomponent is, it's not so much,
hey, here's a new way of doingthings.
Please change it. It's more likefinance people are already
experiencing deep pain andfrustration. Yeah. And we say
(14:25):
you have these things. Let usshow you how you can solve these
things.
Right. That's, you know, on ahigh level, how you sell any
piece of software, but thereality is, like, we're not
inventing these pains. Theyalready exist. We just need to
go solve them. I mean, yeah,that makes a lot
Adam Larson (14:40):
of sense to to just say, hey. This is what's
happening. Here's how you solveit. And a lot of times just
getting in the doors is part ofthe battle. When you're when
you're creating your technologyand and utilizing these these
new ways of hitting those painpoints, You know, how are how
are you using things like AI andstuff like that to kind of
(15:00):
improve and and make it a betterexperience?
Siqi Chen (15:03):
Yeah. We think about AI quite extensively, and, I'm
deep into it. Most peopleTwitter follow me for, AI. So at
a high level, we have a veryspecific philosophy on AI. And
so if you think about like whatmost people's experiences with
AI are, it's usually AI as thisseparate, almost creature apart
(15:24):
from you, right?
The chat interface you'rechatting with chat GPT. And, you
know, there's a lot of hypoagents and agents are like the
separate independent entity.That's like going to do a bunch
of things for you. That is notgenerally how we think a good
expressions of AI should be. Ifyou think about the most
(15:45):
economically, productive usecase valuable use cases of AI
today, it's for softwareengineers and the way they use
AI is very different from chatGBT, copilot cursor and things
like that.
What you do is you're writingcode and AI is looking, watching
you writing code and say, oh, isthis what you mean? And it's
going to like help you movefaster and auto complete and do
(16:06):
things like that and make yourcode more understandable, help
you write faster as you'reworking. You don't have to chat
with it. You don't have to askquestions because the problem
with chat is that most peopleactually don't really understand
what AI can do. And if you don'tknow what AI can do, you don't
know what to ask it to do.
And it doesn't do anything untilhe asks you to do something. And
(16:27):
when it does do something, itdoesn't know a whole lot about
your intent, who you are, whatthe business is. And so with,
with runway, we think about itreally differently. So we try to
have AI do work for you withoutyou having to ask it. And it
knows all the things that are inyour model and the context of
your business to help you workfaster.
(16:47):
So one really simple featurethat we have for example, is
when you're, when you create arow or a formula or a page or a
model, and you open it up, thereis actually a human, English
description, readable,description of the, how the
formula works. It's just likeexplanation. Oh, for expenses is
the sum of, your head count andhead count is turned by like
(17:10):
people who have a set, a definedrole, and you divided by 12 to
get to your monthly, cash burn.And so like, it'll explain these
things in human, in Englishwithout you happy to ask. Right.
And it doesn't even look likeAI, but is it talk AI on the
back end? So we believe in asort of really understated use
(17:31):
of AI that accelerates work andmakes it more understandable and
accessible for people inside andoutside of finance. So they
don't have to ask you to explainhow the model works.
Adam Larson (17:42):
Well, and and not everybody can become a prompt
engineer to properly ask AI theproper thing. And I think that's
why a lot of people don't alwaysget all they can out of it
because they don't understandwhat prompt engineering is.
That's right.
Siqi Chen (17:54):
I mean, prompt engineering is like basically
managing a junior, perhapsautistic, but also very, very,
very smart employee. Right. Andso it's it's, but like it takes
practice and most people likeassume that either it's very,
very difficult or it doesn't itthey they sort of, like,
(18:17):
underrate it in terms of howpowerful it it can be and how
much it can help. And so if wecan understand what people's
intent is and accelerate thatwork, we can sidestep that that
problem.
Adam Larson (18:30):
And I think that's that sounds amazing. And in
perfect practice, I'm sure itwere Roarke's amazing, but not
everybody's ready for that typeof action within their finance
team. I don't I don't know
Siqi Chen (18:43):
that everybody's ready to do that. Yeah. I mean,
I think that the great thing isyou don't have to be. Right?
Like, you know, if you don'twanna use it, and that's totally
fine.
But that's the real differencebetween like how we're thinking
about and how we'd express AIrunway versus a channel
interface. Right. You don't haveto do anything. You can just do
it and it'll support you bymaking things more
(19:05):
understandable without youhaving to ask or accelerate your
work without you on the ass inthe same way that cursor or like
you have copilot does. And the,the way we know another mental
model we have is the re whatwe're building is a tool for
thinking, Right.
And that's a very different toolthan what exists in the
marketplace in the space. Likeyou have tools that, that
(19:27):
automate, right? That likeconnect to your data sources and
help you actualize your model.You have tools that are workflow
tools, which help you make iteasier for you to get
information and having peoplefill in budgets and improve all
of those things. But whatdoesn't exist and what we're
building around, right?
That's fairly unique. There,there aren't really good tools
(19:47):
for thinking outside of Excel.And if you are not a very good
tool for thinking, people willchoose to put things on a
spreadsheet. And that's how youend up with 40 different
versions of that spreadsheet.Yeah.
But because we're building atool for thinking, what that
means is we use AI to help youthink, not do the thinking for
you. And that's the differencebetween sort of an agent and a
creature and versus somethingthat is a copilot that helps you
(20:10):
work. Mhmm.
Adam Larson (20:11):
So can we talk a little bit more about that,
about a tool that helps youthink? Because Yeah. I think a
lot of times people feel like,oh, the AI tool is gonna, you
know, gonna make my job easier,but I feel like sometimes we use
it in a way that it it causes usnot to think as much. And so
Yeah.
Siqi Chen (20:28):
Can we talk a little bit about that? Absolutely. I
think that's a super importantpoint, which is why we didn't
start. I understand that atechnical deep technical level,
like, how AI works and all allthis stuff. But because of that,
we did a start with AI.
Because if you think about whatour tools that help us think
(20:51):
they are tools that one help usoffload ideas from our head in a
way that it makes sense to us.Right. So the, the, the example
that I have is when you thinkabout as an operator or a
product person or businessperson, the business, you think
about product roadmaps. Youthink about marketing plans. You
think about hiring plans,fundraising plans, and you write
(21:13):
them down and it lives in notionor somewhere or Google docs as a
document.
Yeah. You probably don't thinkabout the business in the form
of a wall of numbers. Right? So,so the format of how the model
and finance presented justdoesn't match math to what is in
people's heads. And so part of abetter tool for thought is can
(21:37):
you get it closer to that?
And so we had to invent thingslike plans, and you can see it
on our, on our site, right? Likeyou can now offload a roadmap
into runway and you can say,we're gonna have this marketing
campaign and we're gonna improvethe product in this way with
this feature. And that's beenimproved conversion by X. And it
(21:58):
looks like a roadmap. It lookslike a Gantt chart.
You can move it left and rightin tiny. You can move it up and
down. Right. And what thatreplaces is, you know, a really
simple example is you have yourspreadsheet and let's say
there's a row called conversionrates. And you say it's 25%.
But then you think that in 3months from now, there's some
new sales enablement thing orsome new feature that's gonna go
(22:20):
make that 25%, 35%. So what doyou do? You, you go to a
spreadsheet and you type 35,right? And that's all. And
Because the formula is like,whatever it was last month,
every month after that is gonnabe 35%.
Right? So Yeah. So it's that'sthat's fine. But the problem is
like, all of the context about,like, why did it go from 25 to
35%. Like, what who's gonna doit?
(22:43):
What are you gonna do, is isn'tthere. And so what we observed
is that that this continuityrepresents like something
happened. You've had to dosomething to make that number
change. Mhmm. And we had to hecreated new containers called
plans that lead you connect thatcontext with your model.
And now you have roadmaps andprojects and you just move them
(23:03):
around and now it's a lot moreaccessible. And so there's all
these different things that wethink about, and this is not AI.
This is design, right? Yeah.Yeah.
And that's what, you know, agood tool for thought is. So
it's Excel Excel is a bettertool for thought, because, you
know, you don't want to do allthese calculations in your head
and you're gonna see it. And nowit's a little bit closer to how
you imagine it, but there's likemore things that you can do to
(23:25):
make it even closer, and that'show we think about this. It's
more of a design problem than AIproblem.
Adam Larson (23:29):
Well, yeah. Because, like, in today's world,
let's say you're doing it in aspreadsheet. You might put a
comment on that field or put anotes thing, and you'll see a
note that may or may not makesense, and you also find a
document that that connects to.But what you're describing is
is, like, oh, maybe click thatand boom, you see the plan that
it connects to that shows whatwhat's
Siqi Chen (23:47):
right. Right. And you can group them. Right? So, you
know, if you think about, like,let's say you have a plan for
marketing and this is part ofour standard demo.
Right? Yeah. Mark your marketingplan consists of like, okay,
we're gonna increase likecommission rates or like spend
more marketing and there's like,it affects maybe like 10 rows.
Adam Larson (24:03):
Mhmm.
Siqi Chen (24:03):
Right? But then you say, you go to the board and you
present this plan and then theboard says, Hey, this is a great
plan. What happens when we pushback the spend by a quarter?
You've everyone has done thischanging. The timing of things
in Excel is the worst thing.
Cause you have to like find the,the, the three things that you
would change, move it to 3months later and make the old 3
(24:27):
months the old thing, and thendo it for every other road that
was impacted. But when you haveplans, you literally drag the
entire thing 3 months, andyou're done and have a second.
Right? That's a power of havingobstructions and not closer to
how people think rather than,you know, what the tool is
natively. And so this is notlike a knock on Excel.
Excel is designed to do anythingreally, really quickly being
(24:50):
probably flexible. But businessplanning is a pretty specific
thing And you can build a lot ofmore powerful instruction on the
top that make it a lot moreeasier to understand and faster
to manipulate. So we can so whatyou're saying is basically if
Adam Larson (25:05):
we have the right tools in place, we can be better
strategists within ourorganizations? Totally. I mean,
like that is like the, you know,story of human technology,
right? Like,
Siqi Chen (25:15):
everything like if you think about what I I
considered earliest form ofhuman human technology speech,
right? Like speech allowed us tocoordinate, people within
earshot right now. It's not inour heads anymore. We can tell
people, Hey, here's what we'regonna do. Here's the plan.
And then the next great humantechnology was writing. So
(25:37):
instead of just coordinatepeople with an earshot, now we
could coordinate people with,from great distances and over
great times. Right. Because wecan write down our intents
through laws and culture. Andthe next big I mentioned was
like, okay, now we can like talkto people instantaneously across
the ocean, right?
The Telegraph and email and theinternet. And as in modern days
(26:00):
in, in Mo in the modern day, themost complicated organizations
have, are so interconnected. Andso you have these spreadsheets
that model these differentrelationships, right? And
they're extremely useful becausenow we don't have to do all
these calculations in our head.It'll do the calculations for
us.
And, you know, we see what we'retrying to do a runway as another
(26:24):
step towards that, where, okay,you have these, these tools that
help you do calculations reallyquickly, but it's disconnected
from the level at which businesspeople think about our business.
Mhmm. Romance and plans andhiring. And we can like create
another layer on top of that tomake it even closer. And
ultimately allows us tocoordinate more complex things
(26:44):
with more clarity, faster.
That's how we think about arunway as a tool for thought.
Adam Larson (26:50):
Yeah. Well and and I think, another big part of it
is is is allowing yourorganization, people within your
organization the space to thinkas well. I feel like we're so
busy and there's so many tasksthat we have. We don't allow
ourselves enough space to thinkto be able to implement these
things properly.
Siqi Chen (27:08):
That's right. Yeah. I mean, when there's lack of
clarity or the first thing thatwe reach for is work harder with
more time. But like, you know, Ithink this is a quote from Trey
Zoshi. He was like a, Stripe aproduct leader.
And he has this quick quote,like thinking is cheap. You sh
you should think more. Yeah.Because you know, if you don't,
(27:31):
then you do all this work andsome of them can't be undone and
it can be very difficult to undoand it's gonna get more
extensive later. So think more.
Which is easy enough to say, butlike, how do you think more if
you don't have the understandingof the content and the context
that, reflects the reality andthe breadth of the entire
(27:55):
business. Right. It's sointerconnected. And the way we
try to solve for that problemtoday is like everyone is in
their own silo. Just focus onyour thing and, you know, smart
people will figure it out.
That doesn't work in largeorganizations. Right. People
make so many decisions andmisalignment happens is because
not everyone has the sameunderstanding. And it's our
(28:18):
belief that it's actuallypossible to give people a lot
more of that context, like whatthe trade offs are, like why the
company chose plan a instead ofb, what the trade offs are, that
was thought through in astrategy, if you give them
better tools.
Adam Larson (28:34):
What I'd even argue too, it's even harder in smaller
organizations because your silosare one person, and sometimes
it's hard to break across thosesilos because that one person is
doing so many things. Theybarely have time to think, and
so you can't collaborate aswell. And so we have to break
down those walls and and getcommunication going.
Siqi Chen (28:51):
Yeah. So much of good communication is offloading.
What's in people's headsexternally. Right? Like that's
why, that's why people lovewhiteboards.
Right. It's like, it's a sharedcanvas where all kids are
offloaded thoughts andmanipulating quickly. Right. On
some level informationtechnology are different forms
of this, the performance ofpeople offloading their thoughts
(29:12):
and collaborating with people tochange things and see what
happens. And yeah, like inrunway, we think about that
pretty explicitly.
And again, like our main insightis that the way the tools that
you have to offload thesethoughts don't match the
thoughts. You have a roadmapthing. It's very difficult to
offload that roadmap into aspreadsheet and especially as
(29:33):
necessary to understand theconsequences of it, but it right
now can contain the plan. Sothen you put the plans on a
document, but a document can dothe calculations. So
Adam Larson (29:45):
it's a very interesting and fascinating
impactful design problem. Itreally is. Because, you know,
you have, like, especially inorganizations, you know, may you
might have the senior leaderswho have written all their plans
out in a word document, and thenthe CFO puts that into a plan in
a in a spreadsheet, but there'snothing that ever connects those
things together necessarily.Yes. That is exactly one of the
(30:05):
core innovations that we'vecreated here.
Wow. I mean, I think that'samazing, and I think more
organizations should reallythink about how do we break down
these silos in a bigger way.Because a lot of times we talk
about breaking down silos fromleadership perspective, and it's
a lot of theories, but how do wepractically actually do it so
that people are kind of forcedto work together based on
Siqi Chen (30:24):
what you're saying? Yeah. I mean, there's 2 ways to
do it, and one way is notsustainable. One way is like,
you can just throw money at theproblem actually. And a lot
large companies tend to do this.
So that is basically what sortof the role of the embedded
finance partner in thedepartment does. Yeah. It's like
embedded in a project departmentand it's able to your is this
(30:46):
person who understands theoverall picture of the business
and is able to translate that.And that's your entire job.
Right.
It's just basic communicationand this translation layer. And
I think that's actually like areally good pattern, at larger
companies. Like, like the effortis good. But like, you know, my
belief is that a lot of this isbecause of lack of good tooling.
(31:09):
If the tools do not make itunderstandable for people to
use, do not make the modelitself understandable.
Don't make it easy for you tolike play with different things
and understand the impact of it.You're still going, going to
have to go through a person todo it. And great tools are
ultimately, I think what unlockssilos, I think it's really
underrated in terms of how muchtools impact human behavior. I
(31:33):
mentioned Sigma earlier. Anothergreat example is Amplitude,
right?
Amplitude is analytics platform.So I was a seed investor in our
company and I invested becauseat Zynga, I was head of product
and we had a data department andthe way you would get your data
(31:55):
is you would email your dataanalyst team and with your query
or what you wanna see, and theywould go in and write a report
in SQL format it and email itback to you a day later. Right.
And that's, yeah, that's how itworked. Yep.
But what the new generation ofproducts like amplitude and mix
panel has enabled is that nowyou can just do it yourself,
(32:16):
right. Very, very quickly. Andthey built obviously incredibly
successful, you know, multi$1,000,000,000 businesses that
way. But what is underwriting ishow much is it's changed the
norms of teams. So now we kindof take for granted that anyone
in company who works at like acompany with small and large
just gets to know how manyusers, how many new users, what
(32:40):
retention is, what engagementis.
Right. That's just taking forgranted now. But we forget that
like before these tools that waslocked down. Yeah. Not everyone
got got to see that you had togo through the data team and
there was security, but noweveryone sees it and we, we
think, oh, that's on thebalance, a really good thing.
Cause people understand wherethe product is going. There's no
(33:01):
equivalent for finance. We justtake it for granted. It's like
under, it's like super lock andkey. And part of it is because
the tools just aren't designedfor that use case.
Right. A really, you know, Imentioned a really simple
example where if you wanted todo that and your model is on
Excel, you physically cannot doit. The reason why you can't do
(33:22):
it is because you can't hide thesingle column called salaries in
Excel or Google sheets, becauseall calculations are technically
done on the front end. Like,it's architect from the bottom
up to not enable that, to notmake that possible. So it means
you have to have somethingdifferent from the ground up if
you want to enable that usecase.
Adam Larson (33:40):
I really do. Man, Siggi, this has been a awesome
conversation. I just reallywanna thank you for coming on
the podcast, for sharingeverything you're doing at
Runway and all the, differentideas, and I really hope
everybody checks it out andreally starts thinking thinking
more. Let's start thinking more.
Siqi Chen (33:55):
Let's start thinking more. Thank you, Adam.
Announcer (34:00):
This has been Count Me In, IMA's podcast providing
you with the latest perspectivesof thought leaders from the
accounting and financeprofession. If you like what you
heard and you'd like to becounted in for more relevant
accounting and financeeducation, visit IMA's website
at www.ima net.org.
Welcome to Count Me In. In this episode, we are
joined by the innovative SiqiChen, CEO of Runway, and a 4
time CEO and 3 time founder.With a rich background in math,
technology, and productmanagement, Zeki shares
groundbreaking insight intoleveraging AI for enhanced work
efficiency and intuitivethinking tools. We discussed the
nuances of prompt engineering,Runway's revolutionary approach
(00:43):
to AI, and the importance ofdesign and financial planning.
Zeki contrast traditional toolslike Excel with cutting edge
solutions, aiming to democratizefinancial data access and
improve collaboration acrossorganizations.
Tune in for an engagingconversation that promises to
reshape how you think about AI,financial modeling, and the
future of work. Well, Siqi, I'mreally excited to have you on
(01:11):
the podcast, and we're gonna betalking a lot about, you know,
FP and a and financial planningwithin organizations. You know?
But I wanna talk a little bitabout your story because you
don't have the normal, you know,starting or running an
organization, a financialorganization, story. You're not
a typical, hey.
I started and did my, you know,did my accounting with Deloitte
(01:33):
or KPMG and, you know, kindawent up that way. You don't have
that typical story. So I wantedto talk a little bit about your
story first.
Siqi Chen (01:39):
Yeah. Happy to share. You're right. This is my first
time running a finance org andbuilding anything in B2B much
less, in finance software. Mybackground is in math.
I majored in math, in fact, andit's mostly in technology. So,
my first job, during college wasworking on the Mars rovers at
(02:02):
JPL, in machine vision. And overthe course of my career, now a 4
time CEO, 3 time founder. And Inever really felt comfortable
with finance to the extent thatI was asking our 1st head of
finance, whether I should justgo get an MBA. And he told me
no.
(02:23):
Do it. So that very muchdirectly inspired what we're,
what I'm doing now. But, youknow, as in retrospect though, a
lot of the things that I workedon, in my background is in
product and growth eventually.Right? So I ran product at a
company called Zynga.
It was, it made Farmville, whichmany people may have know and
(02:43):
now a public company. But atthat company, product management
was about modeling. It's aboutmodeling retention and how users
come in and how we monetizethem. And we did it in
spreadsheets and that reallyinformed how we think about
finance at runway. Because whatI realized is that, you know,
(03:04):
from a first principleperspective, financial models
are not actually different than,these product growth models.
They're in different units andthe terminology is different.
But at the end of day, what allof these models do is they are
simulations of something. Theylet you play forward, you know,
time travel into the future tosee what's gonna happen to the
(03:24):
business today and the decisionsthat you make today. And so I
very much, take a broader firstprinciples view of what finance
is about. I think that's beeninteresting for the product and
for our organization.
Adam Larson (03:36):
That really has. It's very it's a unique
perspective, but I think withinorganizations and with the
finance organization,especially, you kinda need to
take a different approach to,first of all, get new clients
and and make it more applicable.But how do you take, like,
things like finance with allthese different lingos and make
it understandable for, you know,finance professionals, but also
other people within the companyor even new clients? Yeah.
(03:59):
That's a great question.
One of our core beliefs is thatpools inform human behavior more
than people give credit, givethem credit for. Okay. And so
we, right now, you know, we, welive in a world where finance in
most organizations is verysiloed
Siqi Chen (04:16):
from the rest of the departments, right? It feels
like black box and thisfrustrates, not just the rest of
the company, but also thefinance leaders, right? Finance
leaders don't feel appreciatedor understood for the value that
they bring to the table, whichis substantial. And a lot of
this is lingo and it isunderstanding that it's
technical, but it's also theavailability of tooling to make
(04:39):
it accessible andunderstandable. So by way of a
simple example, let's say yourmodel lives on Google sheets or
Excel.
And you are a very forwardlooking finance person, and you
want everyone in the company tounderstand how the business
works. So you want people, youwant to share this model with
the rest of the company. Youdon't. And people think it's
because the finance leadersdon't want to. But the reason
(05:02):
why they don't want to isthey're afraid of few really
important invalid things.
The first thing is you giveeveryone access and people might
accidentally change somethingwithout, you know, that's like
legitimate. You know, you wantto have those locked in
operating plan and so am I fatfinger or something? The second
reason is there is no capabilityon Google sheets or Excel to
hide a single column. And thatcolumn is usually called
(05:24):
salaries. Right.
Because all calculation isarchitected to be done on the
front end. And so because oftools don't allow you to do it,
people don't do it. And then thefinance department and leaders
think, oh, they, they, they getthought of as, as like this,
this very protective departmentand people who are very
sensitive to who gets to seewhat, when in actuality the
(05:46):
tools and support it. And if youlook at sort of the last
generation of software fordesign and analytics, right?
Like Figma is probably a goodexample for designers before the
invention of Figma.
It was rude to look over theshoulder of a designer as they
were working. Right. But afterFigma happened because it was so
(06:10):
collaborative of so accessible,Everyone lives in Sigma, whether
you're a product person or abusiness person or engineer. And
it's okay because the productwas designed to do that. There
is no such product for financetoday.
And we think that's a hugeopportunity. And it can change
behavior. And the behavior iswhat it is because the products
don't do that. Wow.
Adam Larson (06:35):
It's like creating Figma for finance, basically
what you're saying, right?
Siqi Chen (06:40):
That is literally the tagline for our seed pitch. And
it was inspiration for why thecompany existed. In 2020, I was
running a company called sandboxVR, very complicated retail VR
hardware. I was justdisappointed CEO 6 months ago
and COVID hits and we were doingscenario planning, and very
(07:01):
quickly our revenue went from,you know, tens of 1,000,000 to
0. And we had to figure out howwe're going to survive.
We had like 400 employees, wentfrom 400 to 10 and I laid myself
off and retired executive team.And 30 minutes afterwards, I
talked to our CFO, our at CFOfrom sandbox. And I asked him,
Hey, we have really amazingtools today. We have Figma and
(07:24):
notion and air table. Surelythere was something better we
could have used to do all thescenario planning than Google
sheets.
We just didn't bother using it.And he laughed and said, no,
this is like all there is. So Italked to Andreessen team and
that's how wrong we got started.So that's actually the direct,
inspiration. And Dylan, thefounder of Figma is one of our
first investors for the samereason.
Adam Larson (07:45):
Wow.
So how is that gonna, how canthat revolutionize an
organization if you can have figa Figma for finance and
accounting team? Because a lotof times, you know, you have
these big old spreadsheets, likeyou said, Google Sheets. You
have, you know, these, ERPsystems like that. They think
they can they can do similarthings, but not exactly how
you're
Siqi Chen (08:03):
describing. Yeah. So, I think it's probably one of the
most impactful things someonecan work on for business. So one
of the, my observations is youspend 100 of 1,000 of dollars,
like implementing something likea finance platform or ERP, and
(08:25):
you pay a 100, a few $100,000.And on top of that, you pay
maybe an extra, a few 100,000hours to implement it over the
course of 6 months or 2 years,depending on the company.
And what happens is you spendall this time and you connect
all this data and you have thisforecast, but the minute the
business changes, you have tolike pay somebody again to
(08:46):
change it because you can'ttouch it. It's so delicate and
sensitive and fragile. And theother thing that happens is
let's say a department leader,like a product person or a GM or
whoever has some idea about whatthey're gonna do in the business
and they need to model it. Thelast thing they're going to
going to go to try that is goingto start with Anaplan or
(09:08):
something like that. The firstthing that you reach for is the
most convenient modeling toolthat's been ever invented, which
is Excel.
Right? And that seed of an ideathen becomes a real model. And
eventually you have aspreadsheet model in every
different department. And soyou've spent 100 and 1,000
dollars for this thing, but thereal reality, the real model of
(09:32):
how these different departmentswork lives completely outside of
that. And as a result, peoplehave misalignment and there's no
visibility.
And your great model in your FPand a platform doesn't reflect
any of that. And so if you canbuild something that is people's
first choice to think of anidea, because it's easier,
(09:54):
faster, and more intuitive, theimpact is that everyone in the
company actually gets to have areal view of the business. A
true view of the business is upto date. So when I look at the
product, you know, the, theproduct ropes part of the model
that actually represents whatthe current thinking is of the
product team, what the marketingplan is, what the sales plan is
(10:15):
instead of like sort of a, maybea 3 month delay to view entered
because I sent a spreadsheetover slack to the sales head to,
to fill it out and then paste itback in. And the the dream that
I think we're very close toachieving.
And in some of our customersalready achieved this as imagine
(10:37):
a designer or engineer or a PMbe able to say, if we work on X
instead of Y here is how it'sgoing to impact margin and
growth for our company 2 yearsfrom now.
Adam Larson (10:48):
Wow. That conversation does not happen,
which is actually wild becausewe all are working on the
business. Everything that we doimpacts the bottom line
somewhere. That is a possibleworld. And it's not possible
today because the tools simplyaren't there, but it's totally
achievable and totally billable.
Well, yeah, it seems like in intoday's world, you know, if
(11:09):
you're budgeting or or doingthese these big old strategic
planning things, it's thismonths months long processes.
You're doing spreadsheets.You're throwing things around.
Then you have meetings with thebig wigs, and then the the big
wigs talk to the peopleunderneath them, and then it's
spread out to everybody. Andit's this long process, but what
you're describing sounds like itcould be in a matter of days.
You're like, hey. This is whatwe're planning on. This is what
(11:30):
we're looking to do. And and itbreaks down the silos that are
today built up in organizations.
Siqi Chen (11:35):
That's right. I mean, we have some of our customers
integrating Jira into Runway,which, you know, is unheard of
for a finance platform. Wow. Butto your point, it's not just
about the annual quarterlybudgeting. Right?
When you're thinking about thenext few items on your roadmap,
where your next project is, youshould have a conversation and
you should have understanding oflike, what is this going to move
(11:55):
and how does it impact overallbusiness? And so one of the big
opportunities is if you thinkabout how finance has evolved as
a discipline, right? I think weall know it's evolving from this
sort of backwards lookingcontrols reporting to more of a
forward looking decision makingstrategic role. What does, what
(12:17):
does it mean to be strategic?Right?
Like it means you have to lookinto the future, but also it's
not just a forecasting thefuture accurately. It's a
understanding the causality ofwhat are the things that you can
do today to move the future.And, and that's underappreciated
and, and you need better toolsfor everyone to have that same
(12:38):
understanding here are the keydrivers that drive growth, not
just like we're good forecasting10% year over year growth or 20%
year over year growth. Anyonecan do that. Right?
But, like, it doesn't help youmake better decisions. And,
ultimately, that's whatstrategic finance is really
about, helping the entireorganization make better
decisions. So how
Adam Larson (12:56):
do you how do you talk have that conversation,
especially with people who'vebeen in within the finance
function, the accountingfunction in organizations for
years? They've always done itthe same way, and a lot of times
change comes to the finance teamlast because there's so much
that they're doing. It's hard toimplement these other things.
How do you have thatconversation to to kind of say,
hey. You can do something new?
You, I mean, yeah, our our ourmain move is to show them the
(13:19):
product. Okay.
Siqi Chen (13:21):
So if you look at our website, if you click our
videos, you know, it's verydifferent from a typical B2B
enterprise, like type ofwebsite. Yeah. Or you just have
like smiling people, likeshaking hands from stock photos.
The software, like I'm a productperson. Software is designed to
be used.
Of course. Like you have to seeit. It has to be good because
(13:43):
if, if you're not showing offthe product, you're not showing
people how it can work, thenyou're gonna buy software that
is gonna sound good on paperbecause it's really well sold,
but it actually doesn't do thething. And that's our
observation. People have beentalking about connected finance
and strategic planning for like10, 20 years.
It doesn't do the thing. And youhave to invent new product
(14:04):
abstractions and experiences inorder for it to do the thing.
And it's a much higher bar. Sothat's like one component is how
we knew it. But the othercomponent is, it's not so much,
hey, here's a new way of doingthings.
Please change it. It's more likefinance people are already
experiencing deep pain andfrustration. Yeah. And we say
(14:25):
you have these things. Let usshow you how you can solve these
things.
Right. That's, you know, on ahigh level, how you sell any
piece of software, but thereality is, like, we're not
inventing these pains. Theyalready exist. We just need to
go solve them. I mean, yeah,that makes a lot
Adam Larson (14:40):
of sense to to just say, hey. This is what's
happening. Here's how you solveit. And a lot of times just
getting in the doors is part ofthe battle. When you're when
you're creating your technologyand and utilizing these these
new ways of hitting those painpoints, You know, how are how
are you using things like AI andstuff like that to kind of
(15:00):
improve and and make it a betterexperience?
Siqi Chen (15:03):
Yeah. We think about AI quite extensively, and, I'm
deep into it. Most peopleTwitter follow me for, AI. So at
a high level, we have a veryspecific philosophy on AI. And
so if you think about like whatmost people's experiences with
AI are, it's usually AI as thisseparate, almost creature apart
(15:24):
from you, right?
The chat interface you'rechatting with chat GPT. And, you
know, there's a lot of hypoagents and agents are like the
separate independent entity.That's like going to do a bunch
of things for you. That is notgenerally how we think a good
expressions of AI should be. Ifyou think about the most
(15:45):
economically, productive usecase valuable use cases of AI
today, it's for softwareengineers and the way they use
AI is very different from chatGBT, copilot cursor and things
like that.
What you do is you're writingcode and AI is looking, watching
you writing code and say, oh, isthis what you mean? And it's
going to like help you movefaster and auto complete and do
(16:06):
things like that and make yourcode more understandable, help
you write faster as you'reworking. You don't have to chat
with it. You don't have to askquestions because the problem
with chat is that most peopleactually don't really understand
what AI can do. And if you don'tknow what AI can do, you don't
know what to ask it to do.
And it doesn't do anything untilhe asks you to do something. And
(16:27):
when it does do something, itdoesn't know a whole lot about
your intent, who you are, whatthe business is. And so with,
with runway, we think about itreally differently. So we try to
have AI do work for you withoutyou having to ask it. And it
knows all the things that are inyour model and the context of
your business to help you workfaster.
(16:47):
So one really simple featurethat we have for example, is
when you're, when you create arow or a formula or a page or a
model, and you open it up, thereis actually a human, English
description, readable,description of the, how the
formula works. It's just likeexplanation. Oh, for expenses is
the sum of, your head count andhead count is turned by like
(17:10):
people who have a set, a definedrole, and you divided by 12 to
get to your monthly, cash burn.And so like, it'll explain these
things in human, in Englishwithout you happy to ask. Right.
And it doesn't even look likeAI, but is it talk AI on the
back end? So we believe in asort of really understated use
(17:31):
of AI that accelerates work andmakes it more understandable and
accessible for people inside andoutside of finance. So they
don't have to ask you to explainhow the model works.
Adam Larson (17:42):
Well, and and not everybody can become a prompt
engineer to properly ask AI theproper thing. And I think that's
why a lot of people don't alwaysget all they can out of it
because they don't understandwhat prompt engineering is.
That's right.
Siqi Chen (17:54):
I mean, prompt engineering is like basically
managing a junior, perhapsautistic, but also very, very,
very smart employee. Right. Andso it's it's, but like it takes
practice and most people likeassume that either it's very,
very difficult or it doesn't itthey they sort of, like,
(18:17):
underrate it in terms of howpowerful it it can be and how
much it can help. And so if wecan understand what people's
intent is and accelerate thatwork, we can sidestep that that
problem.
Adam Larson (18:30):
And I think that's that sounds amazing. And in
perfect practice, I'm sure itwere Roarke's amazing, but not
everybody's ready for that typeof action within their finance
team. I don't I don't know
Siqi Chen (18:43):
that everybody's ready to do that. Yeah. I mean,
I think that the great thing isyou don't have to be. Right?
Like, you know, if you don'twanna use it, and that's totally
fine.
But that's the real differencebetween like how we're thinking
about and how we'd express AIrunway versus a channel
interface. Right. You don't haveto do anything. You can just do
it and it'll support you bymaking things more
(19:05):
understandable without youhaving to ask or accelerate your
work without you on the ass inthe same way that cursor or like
you have copilot does. And the,the way we know another mental
model we have is the re whatwe're building is a tool for
thinking, Right.
And that's a very different toolthan what exists in the
marketplace in the space. Likeyou have tools that, that
(19:27):
automate, right? That likeconnect to your data sources and
help you actualize your model.You have tools that are workflow
tools, which help you make iteasier for you to get
information and having peoplefill in budgets and improve all
of those things. But whatdoesn't exist and what we're
building around, right?
That's fairly unique. There,there aren't really good tools
(19:47):
for thinking outside of Excel.And if you are not a very good
tool for thinking, people willchoose to put things on a
spreadsheet. And that's how youend up with 40 different
versions of that spreadsheet.Yeah.
But because we're building atool for thinking, what that
means is we use AI to help youthink, not do the thinking for
you. And that's the differencebetween sort of an agent and a
creature and versus somethingthat is a copilot that helps you
(20:10):
work. Mhmm.
Adam Larson (20:11):
So can we talk a little bit more about that,
about a tool that helps youthink? Because Yeah. I think a
lot of times people feel like,oh, the AI tool is gonna, you
know, gonna make my job easier,but I feel like sometimes we use
it in a way that it it causes usnot to think as much. And so
Yeah.
Siqi Chen (20:28):
Can we talk a little bit about that? Absolutely. I
think that's a super importantpoint, which is why we didn't
start. I understand that atechnical deep technical level,
like, how AI works and all allthis stuff. But because of that,
we did a start with AI.
Because if you think about whatour tools that help us think
(20:51):
they are tools that one help usoffload ideas from our head in a
way that it makes sense to us.Right. So the, the, the example
that I have is when you thinkabout as an operator or a
product person or businessperson, the business, you think
about product roadmaps. Youthink about marketing plans. You
think about hiring plans,fundraising plans, and you write
(21:13):
them down and it lives in notionor somewhere or Google docs as a
document.
Yeah. You probably don't thinkabout the business in the form
of a wall of numbers. Right? So,so the format of how the model
and finance presented justdoesn't match math to what is in
people's heads. And so part of abetter tool for thought is can
(21:37):
you get it closer to that?
And so we had to invent thingslike plans, and you can see it
on our, on our site, right? Likeyou can now offload a roadmap
into runway and you can say,we're gonna have this marketing
campaign and we're gonna improvethe product in this way with
this feature. And that's beenimproved conversion by X. And it
(21:58):
looks like a roadmap. It lookslike a Gantt chart.
You can move it left and rightin tiny. You can move it up and
down. Right. And what thatreplaces is, you know, a really
simple example is you have yourspreadsheet and let's say
there's a row called conversionrates. And you say it's 25%.
But then you think that in 3months from now, there's some
new sales enablement thing orsome new feature that's gonna go
(22:20):
make that 25%, 35%. So what doyou do? You, you go to a
spreadsheet and you type 35,right? And that's all. And
Because the formula is like,whatever it was last month,
every month after that is gonnabe 35%.
Right? So Yeah. So it's that'sthat's fine. But the problem is
like, all of the context about,like, why did it go from 25 to
35%. Like, what who's gonna doit?
(22:43):
What are you gonna do, is isn'tthere. And so what we observed
is that that this continuityrepresents like something
happened. You've had to dosomething to make that number
change. Mhmm. And we had to hecreated new containers called
plans that lead you connect thatcontext with your model.
And now you have roadmaps andprojects and you just move them
(23:03):
around and now it's a lot moreaccessible. And so there's all
these different things that wethink about, and this is not AI.
This is design, right? Yeah.Yeah.
And that's what, you know, agood tool for thought is. So
it's Excel Excel is a bettertool for thought, because, you
know, you don't want to do allthese calculations in your head
and you're gonna see it. And nowit's a little bit closer to how
you imagine it, but there's likemore things that you can do to
(23:25):
make it even closer, and that'show we think about this. It's
more of a design problem than AIproblem.
Adam Larson (23:29):
Well, yeah. Because, like, in today's world,
let's say you're doing it in aspreadsheet. You might put a
comment on that field or put anotes thing, and you'll see a
note that may or may not makesense, and you also find a
document that that connects to.But what you're describing is
is, like, oh, maybe click thatand boom, you see the plan that
it connects to that shows whatwhat's
Siqi Chen (23:47):
right. Right. And you can group them. Right? So, you
know, if you think about, like,let's say you have a plan for
marketing and this is part ofour standard demo.
Right? Yeah. Mark your marketingplan consists of like, okay,
we're gonna increase likecommission rates or like spend
more marketing and there's like,it affects maybe like 10 rows.
Adam Larson (24:03):
Mhmm.
Siqi Chen (24:03):
Right? But then you say, you go to the board and you
present this plan and then theboard says, Hey, this is a great
plan. What happens when we pushback the spend by a quarter?
You've everyone has done thischanging. The timing of things
in Excel is the worst thing.
Cause you have to like find the,the, the three things that you
would change, move it to 3months later and make the old 3
(24:27):
months the old thing, and thendo it for every other road that
was impacted. But when you haveplans, you literally drag the
entire thing 3 months, andyou're done and have a second.
Right? That's a power of havingobstructions and not closer to
how people think rather than,you know, what the tool is
natively. And so this is notlike a knock on Excel.
Excel is designed to do anythingreally, really quickly being
(24:50):
probably flexible. But businessplanning is a pretty specific
thing And you can build a lot ofmore powerful instruction on the
top that make it a lot moreeasier to understand and faster
to manipulate. So we can so whatyou're saying is basically if
Adam Larson (25:05):
we have the right tools in place, we can be better
strategists within ourorganizations? Totally. I mean,
like that is like the, you know,story of human technology,
right? Like,
Siqi Chen (25:15):
everything like if you think about what I I
considered earliest form ofhuman human technology speech,
right? Like speech allowed us tocoordinate, people within
earshot right now. It's not inour heads anymore. We can tell
people, Hey, here's what we'regonna do. Here's the plan.
And then the next great humantechnology was writing. So
(25:37):
instead of just coordinatepeople with an earshot, now we
could coordinate people with,from great distances and over
great times. Right. Because wecan write down our intents
through laws and culture. Andthe next big I mentioned was
like, okay, now we can like talkto people instantaneously across
the ocean, right?
The Telegraph and email and theinternet. And as in modern days
(26:00):
in, in Mo in the modern day, themost complicated organizations
have, are so interconnected. Andso you have these spreadsheets
that model these differentrelationships, right? And
they're extremely useful becausenow we don't have to do all
these calculations in our head.It'll do the calculations for
us.
And, you know, we see what we'retrying to do a runway as another
(26:24):
step towards that, where, okay,you have these, these tools that
help you do calculations reallyquickly, but it's disconnected
from the level at which businesspeople think about our business.
Mhmm. Romance and plans andhiring. And we can like create
another layer on top of that tomake it even closer. And
ultimately allows us tocoordinate more complex things
(26:44):
with more clarity, faster.
That's how we think about arunway as a tool for thought.
Adam Larson (26:50):
Yeah. Well and and I think, another big part of it
is is is allowing yourorganization, people within your
organization the space to thinkas well. I feel like we're so
busy and there's so many tasksthat we have. We don't allow
ourselves enough space to thinkto be able to implement these
things properly.
Siqi Chen (27:08):
That's right. Yeah. I mean, when there's lack of
clarity or the first thing thatwe reach for is work harder with
more time. But like, you know, Ithink this is a quote from Trey
Zoshi. He was like a, Stripe aproduct leader.
And he has this quick quote,like thinking is cheap. You sh
you should think more. Yeah.Because you know, if you don't,
(27:31):
then you do all this work andsome of them can't be undone and
it can be very difficult to undoand it's gonna get more
extensive later. So think more.
Which is easy enough to say, butlike, how do you think more if
you don't have the understandingof the content and the context
that, reflects the reality andthe breadth of the entire
(27:55):
business. Right. It's sointerconnected. And the way we
try to solve for that problemtoday is like everyone is in
their own silo. Just focus onyour thing and, you know, smart
people will figure it out.
That doesn't work in largeorganizations. Right. People
make so many decisions andmisalignment happens is because
not everyone has the sameunderstanding. And it's our
(28:18):
belief that it's actuallypossible to give people a lot
more of that context, like whatthe trade offs are, like why the
company chose plan a instead ofb, what the trade offs are, that
was thought through in astrategy, if you give them
better tools.
Adam Larson (28:34):
What I'd even argue too, it's even harder in smaller
organizations because your silosare one person, and sometimes
it's hard to break across thosesilos because that one person is
doing so many things. Theybarely have time to think, and
so you can't collaborate aswell. And so we have to break
down those walls and and getcommunication going.
Siqi Chen (28:51):
Yeah. So much of good communication is offloading.
What's in people's headsexternally. Right? Like that's
why, that's why people lovewhiteboards.
Right. It's like, it's a sharedcanvas where all kids are
offloaded thoughts andmanipulating quickly. Right. On
some level informationtechnology are different forms
of this, the performance ofpeople offloading their thoughts
(29:12):
and collaborating with people tochange things and see what
happens. And yeah, like inrunway, we think about that
pretty explicitly.
And again, like our main insightis that the way the tools that
you have to offload thesethoughts don't match the
thoughts. You have a roadmapthing. It's very difficult to
offload that roadmap into aspreadsheet and especially as
(29:33):
necessary to understand theconsequences of it, but it right
now can contain the plan. Sothen you put the plans on a
document, but a document can dothe calculations. So
Adam Larson (29:45):
it's a very interesting and fascinating
impactful design problem. Itreally is. Because, you know,
you have, like, especially inorganizations, you know, may you
might have the senior leaderswho have written all their plans
out in a word document, and thenthe CFO puts that into a plan in
a in a spreadsheet, but there'snothing that ever connects those
things together necessarily.Yes. That is exactly one of the
(30:05):
core innovations that we'vecreated here.
Wow. I mean, I think that'samazing, and I think more
organizations should reallythink about how do we break down
these silos in a bigger way.Because a lot of times we talk
about breaking down silos fromleadership perspective, and it's
a lot of theories, but how do wepractically actually do it so
that people are kind of forcedto work together based on
Siqi Chen (30:24):
what you're saying? Yeah. I mean, there's 2 ways to
do it, and one way is notsustainable. One way is like,
you can just throw money at theproblem actually. And a lot
large companies tend to do this.
So that is basically what sortof the role of the embedded
finance partner in thedepartment does. Yeah. It's like
embedded in a project departmentand it's able to your is this
(30:46):
person who understands theoverall picture of the business
and is able to translate that.And that's your entire job.
Right.
It's just basic communicationand this translation layer. And
I think that's actually like areally good pattern, at larger
companies. Like, like the effortis good. But like, you know, my
belief is that a lot of this isbecause of lack of good tooling.
(31:09):
If the tools do not make itunderstandable for people to
use, do not make the modelitself understandable.
Don't make it easy for you tolike play with different things
and understand the impact of it.You're still going, going to
have to go through a person todo it. And great tools are
ultimately, I think what unlockssilos, I think it's really
underrated in terms of how muchtools impact human behavior. I
(31:33):
mentioned Sigma earlier. Anothergreat example is Amplitude,
right?
Amplitude is analytics platform.So I was a seed investor in our
company and I invested becauseat Zynga, I was head of product
and we had a data department andthe way you would get your data
(31:55):
is you would email your dataanalyst team and with your query
or what you wanna see, and theywould go in and write a report
in SQL format it and email itback to you a day later. Right.
And that's, yeah, that's how itworked. Yep.
But what the new generation ofproducts like amplitude and mix
panel has enabled is that nowyou can just do it yourself,
(32:16):
right. Very, very quickly. Andthey built obviously incredibly
successful, you know, multi$1,000,000,000 businesses that
way. But what is underwriting ishow much is it's changed the
norms of teams. So now we kindof take for granted that anyone
in company who works at like acompany with small and large
just gets to know how manyusers, how many new users, what
(32:40):
retention is, what engagementis.
Right. That's just taking forgranted now. But we forget that
like before these tools that waslocked down. Yeah. Not everyone
got got to see that you had togo through the data team and
there was security, but noweveryone sees it and we, we
think, oh, that's on thebalance, a really good thing.
Cause people understand wherethe product is going. There's no
(33:01):
equivalent for finance. We justtake it for granted. It's like
under, it's like super lock andkey. And part of it is because
the tools just aren't designedfor that use case.
Right. A really, you know, Imentioned a really simple
example where if you wanted todo that and your model is on
Excel, you physically cannot doit. The reason why you can't do
(33:22):
it is because you can't hide thesingle column called salaries in
Excel or Google sheets, becauseall calculations are technically
done on the front end. Like,it's architect from the bottom
up to not enable that, to notmake that possible. So it means
you have to have somethingdifferent from the ground up if
you want to enable that usecase.
Adam Larson (33:40):
I really do. Man, Siggi, this has been a awesome
conversation. I just reallywanna thank you for coming on
the podcast, for sharingeverything you're doing at
Runway and all the, differentideas, and I really hope
everybody checks it out andreally starts thinking thinking
more. Let's start thinking more.
Siqi Chen (33:55):
Let's start thinking more. Thank you, Adam.
Announcer (34:00):
This has been Count Me In, IMA's podcast providing
you with the latest perspectivesof thought leaders from the
accounting and financeprofession. If you like what you
heard and you'd like to becounted in for more relevant
accounting and financeeducation, visit IMA's website
at www.ima net.org.
Popular Podcasts
United States of Kennedy
United States of Kennedy is a podcast about our cultural fascination with the Kennedy dynasty. Every week, hosts Lyra Smith and George Civeris go into one aspect of the Kennedy story.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com