How Procurement Data Predicts Market Trends

Hello, everyone. We're back at Techstrong TV. I am Colleen Coll at Coupa Inspire in Vegas, Las Vegas, by the way.

And my next interview is with Cory Manley. He is the vice president of value advisory and innovation. I wanted to get that right because that is an amazing title, right, Cory?

Thank you. Oh, and thank you for joining us today. I know it's day one of the actual event, so how's it going so far?

Glad to be here. It's going very well. Today's a unique day for us.

Lee, our CEO, talked in the keynote today- Yes ... about the Coupa community now having transacted $10 trillion- Yes ... in our 20 years.

She did mention that, and that's amazing number. I can't even comprehend. It's hard for us to comprehend, too.

Me too. What's amazing about that, that we did not put on main stage, was that today was the day that we actually crossed that line. Oh, well, congratulations.

That's something to celebrate. It is. It's amazing, especially for somebody in your role.

And you want to talk more about your role and how that is connected to that? Yeah, sure. Mm-hmm.

So my team's focused on advisory and research. So we have value and industry advisory, and then Coupa Spend Lab- Yeah ... and a team of data scientists and analysts that support those teams.

When we put together a lens where we look at value, and we also look at value from an industry perspective, along with all of our community data, and the tremendous insights that are hiding in those data- Mm ... that's when we build a very powerful tool that helps us build really important insights for our customers, for the Coupa community, leveraging the data that they create. And that's very important that we're good stewards of those data- Uh-huh ...

and that we're using them to enrich the community. Mm-hmm. And we think about this in two ways.

Yes. One is at the macroeconomic level. This is BSI or business spend index- Oh, yeah ...

that I know you're excited to talk about. Yes. And the other is benchmarking.

Both of which we've done for years. And the benchmarking is at the microeconomic level. Mm-hmm.

These two things come together to really help us understand what's happening in our customers, in market segments, in the economy overall. Yeah. That's nice.

It's fantastic just to hear, especially with the large $10 trillion - It's a mouthful ... worth of, yes, transactional data. So now getting back to the project at- Mm-hmm ...

BSI. Yeah. You built something called Business Spend Index in collaboration with, and I'll get this right, MIT Data Science Lab.

Yeah. Did I get that right? Yes.

Okay. So tell us more about it. People who, like myself, who haven't heard of it, and what is it, and why it's important to a CFO or a CPO.

Yeah. Why should they care? So, BSI is the macroeconomic.

Macroeconomic. It helps us understand what is happening in the economy- Yeah ... and also what might happen in the economy.

Okay. We'll talk about what it is. But first, relevance.

One of the most important reasons for us to leverage these data to understand the macroeconomic, what might happen, and it is predictive- Mm-hmm ... is because it helps businesses make better decisions. Procurement data are the first data that are generated when a business goes from thinking about doing something, talking about it sometimes- Mm-hmm ...

to actually doing the thing. It is the earliest signal that we can capture without reading minds. Okay.

And even if we were reading minds, businesses still wouldn't have spent the money yet. And so this turns procurement into the intelligence hub of the business- Mm ... and the intelligence hub for the customer industries and for the economy.

And it means that the signal created from that first moment where a business goes from thinking about strategy- Mm ... to engaging in an economic transaction gives us a very unique position. Yeah.

And that position is we see it first. Yeah. That means that we can predict, using those data, where the economy might move in different sectors, and that's what BSI is doing.

Well, that's amazing because it's forecasting 90 days ahead. Yeah, with a tremendous forecast accuracy. It's an extraordinary claim, I guess.

We feel the same way, right? " And of course, there's a lot of rigor and validation that happens, and that's why we have academic collaborators because we're very serious about validating the claims that we make and what we learn from our data. Yeah.

We have a very unique culture at Coupa where most businesses are very focused on business problems in the context of business. We're very focused on those business problems, but also applying the academic rigor and lens to those business problems and our data. And that's where serious innovation happens.

That gap between business and academia is not spanned often enough. But when it is and when it's done well- Mm-hmm ... the outcomes can be quite tremendous.

And at Coupa, we have the investment from our leadership to do that. Has there been any experiences or any shifts that you caught that surprised you? Look.

Yeah. Indeed, there have. So a few.

One is around tariffs. Oh, yeah. So in our data, we were able to see the impact of trade shocks.

Mm-hmm. And Liberation Day was a massive trade shock. It also asymmetrically impacted businesses of different sizes.

Mm-hmm. So small and medium-sized businesses were disproportionately affected by the tariff regime. And larger businesses were able to weather it a bit better.

But what happened is businesses of all sizes pulled forward drastically their orders- Mm-hmm ... in anticipation of tariffs coming into effect. Trade then decreased a good bit, and we also saw trade go around the United States.

So countries that normally would trade with the US were trading with each other instead. Mm-hmm. And that is a dramatic shift in the economy.

The asymmetry of the impact is also really important. Larger businesses are better capitalized typically, and- Okay ... and as a result, can hedge a bit more in advance of an expected shift, and that happened- Right ...

by a pretty wide margin. Another one is the conflict in Iran. So- Yeah, that's very current.

It is very current, and it actually caused us to delay the publication of this report- Oh, is that right? because we now have the data- Mm-hmm ... to understand the impact of the conflict in Iran on trade and on the economy.

And when we did have enough data after the start of the conflict to rerun our models, we found some pretty shocking changes in our predictions. Mm-hmm. And so we decided, let's wait a couple of weeks, and let's really make sure that we get this right because it did change dramatically on the BSI end and what we were predicting.

Well, because that's a great example where no matter if you did forecast it, that something always happens where you have to be a little bit more reactive than proactive- Yeah ... in forecasting. That's the thing about forecasting and prediction.

Yeah. It's all fun and games until there's a shock. Yeah.

Right? Until something changes in the system. Mm-hmm.

Tariffs, something changed in the system, and it was dramatic. Unexpected or expected? Unexpected.

Well, in that case, it was signaled, and I think there was- Yes ... a lot of hope that it might not happen. Exactly.

Iran, more of a shock that was not expected at that magnitude. Yes. And so these interrupting actions- Mm-hmm ...

cause shifts that then we need to go back and look at what we were predicting and how those things change. Yeah. Does that affect the world's largest data sets that you have when you have these things happen that goes back in, and it helps with forecasting for later experiences?

Just me trying to find out how this works. Yeah. Yes.

Absolutely. And then how does AI get into this mix? I know that you need a lot of good data for it to work well.

So- Mm-hmm ... how does that work for Coupa's customers? Well, so first of all, the research that we're doing- Mm-hmm ...

especially around prediction, is done using machine learning. Yes. Oh, okay.

These types of things would've been much more difficult and lower resolution just a few years ago. Yeah. Mm-hmm.

And so what we're now able to do is dramatically more detailed, and our goal is to respond to the news cycle with the data when it is happening. Because often we watch the news, we see these events in the news that we understand, or we have the data to understand the underlying fundamentals that are causing the news. Mm-hmm.

And often we see it before it happens, before it's in the news. Yeah. And so we want to get closer and closer to a near real-time understanding of- Hmm ...

the impacts of these events. Are you finding that more difficult, or do you think it'll be easier in the long run? These are engineering and infrastructure challenges, but frankly, we're finding it far less difficult than we expected.

Oh, that's good. Like in the context of Iran, we've been able to very quickly- Oh ... look at this stuff.

Oh, that's fascinating. I love it. So do we.

So based on day one and what you experienced today, is there anything else that if you've been going to any sessions that you might want to talk about that was very insightful, or any hallway track conversations you had? So the thing I love about these events are conversations- Yes ... with customers and prospects.

Same. There's no more valuable conversation because it really helps us understand what customers are experiencing in their industries- Exactly ... and how we can leverage community data to help them make better decisions.

Mm-hmm. What we do when we collaborate with product. Those conversations are invaluable.

Yes. I just had one over lunch that I- Oh ... really appreciated.

And often- Yeah ... we are looking for customers who want to collaborate with us, even in research. Okay.

" Mm-hmm. "Because we want to use them to-"To measure the value that our own program generates. Yeah.

And those are great conversations. We get to collaborate and build future partnerships for research. That's good, and I'm sure that's satisfying for you- Oh ...

Coupa team as well. Extremely. So, my last question, so we will wrap it up.

What does the next generation of data science look like? Where's Spend Labs headed? Yeah.

So, I wish we could predict where data science goes and- But wait, can't you forecast it in nine days? That's something I wish we could forecast. Yeah.

The reality is this, we're looking at the combination of the micro and macroeconomics- Mm-hmm ... so the benchmarking and BSI, the interaction effects between these things, because benchmarks really underlie a lot of what BSI shows us. Mm-hmm.

But also- Yeah ... we're using very high-level summary data to create BSI today. And so, we were even shocked at the results- That's okay ...

given what data we use. Mm-hmm. And so, this is the very tip of the iceberg.

Good. And so, once we publish this report and take about 30 seconds to breathe- ... we have a pipeline of research that we'll take deeper.

We will also, and are already working on more academic collaborations with more labs. We're very interested in academic teams that focus on very specific business process areas- Mm-hmm ... that are in the scope of what we do.

Yeah. Because having that domain expertise from academia, combined with our industry advisory team that is domain expertise from industry and Spend Lab, is a very powerful combination. And those are the pieces that really have to come together for us to build these programs and get these types of insights.

Mm-hmm. There are also, of course, conversations around how we get some of these things into product to help our customers in the Coupa product also make decisions based on the data. There's one more piece, and that is helping customers ask the questions that they don't know to ask.

Well, how do you do that? Great question. You're forecasting, but how you forecast the-- Can you forecast the questions that they don't know how to ask?

Not the questions. Okay. But we do feel pretty, well, for now, but we do feel pretty confident about insights that might lead them down certain decision-making paths.

Mm-hmm. Understanding where certain sectors might go may help us make good decisions about how we spend our capital. And this is something very interesting, of course, to CFOs or to private equity firms or VCs.

BSI looks at a number of industries, and there is a coupling of some of those industries that were not coupled in the past, and largely this is because of the enormous amount of capital outlay in the infrastructure build for AI. Mm-hmm. Trillions of dollars.

And there are other sectors that are not tied to that right now, that are in a very different cycle. And then there are those interrupting actions like Iran or tariff regimes- Yeah ... that also we have to adapt to and help customers understand the impact.

Yeah. That's amazing. Corey, thank you so much.

I enjoyed learning more, especially when it gets very heavy in talking about data science. This was wonderful. Corey Manly, Vice President of Value Advisory and Innovation.

We have more to come at Coupa Inspire, so stay tuned.