Sign up to be notified when new chapters are released.
Ariba: Building a Supplier Network
Founder and Partner, Tidemark
Former SVP/ GM of Supplier Network and Financial Services, Ariba
Ariba was one of the first of a cadre of web-based software companies that set precedents for modern day SaaS titans. The company went public in 1999 as the first B2B Internet company. It started as a procurement application, but by the time it sold to SAP for $4.3B in 2012, Ariba had expanded to spend visibility analysis, sourcing, contract management, invoice management, discovery & marketing, dynamic discounting, and more. At the time of its sale, the Ariba network was the largest global trading network, connecting and automating $300B+ in transactions for 700,000+ companies.
Bob Solomon, Former SVP/GM of Supplier Network and Financial Services at Ariba, is a long-time friend of mine and one of the smartest strategists in B2B platforms. We were fortunate to have Bob sit down with my colleague Phil to talk through the Ariba story.
Their discussion brings Ariba to life. It’s one of the oldest and most complete examples of a supplier extension. In this interview, we’ll discuss:
- Procurement as a control point
- Using a procurement to launch a supplier side wedge
- Building a two-sided supplier network
- Third party tools and network liquidity
- Network pricing and monetization
This interview took place on April 7, 2022. Lightly edited for brevity and clarity.
Phil: Bob, thank you for taking the time today. We are super excited to have you walk us through the Ariba story and your background with the company.
Bob: I started with Ariba in 2000, which many readers will be too young to remember, but was the height of the bubble era. I always tell people that my first day was the highest stock price that Ariba ever reached. I hope it did not drop because I was joining!
When I joined, Ariba was thinking about how to become more vertical. I worked on that for a few years and then I got involved with the Ariba network. You have to remember the company introduced the network about three years after introducing the procurement tool – Ariba was the original “come for the tool, stay for the network” solution. People don’t think about it that way, but that’s what it was. The network was a product that was given away to customers who licensed the procurement tool. And it was also originally given away free to suppliers.
In 2003, once we had gained traction and some volume of transactions on the network, my job was to figure out how to turn the network from a cost center into a profit center and into a product. It was the original many-to-one / one-to-many platform in B2B in some ways.
We hadn’t really fully understood how the network could be monetized. So that’s what I did through the end of my 10-year career at Ariba. It was a really fun ride.
Procurement as a Control Point
Phil: There’s so much to unpack in there, Bob. For our audience who are maybe less familiar with the story, what was the original, key value proposition and catalyst for most customers?
Bob: Remember, we’re in the mid ’90s and people are still buying from catalogs—physical catalogs—then phoning and faxing their orders in. The first insight that Keith Krach and his cofounders had was: Do we need paper orders, phones, fax, email orders? There must be a better way to do this. There are lots of people inside a company who need to buy stuff. They need to look it up. They need to buy it off the right contract. Ariba set out to make it easy to look that up, do workflow around the requisition, route it by the right people, and turn it into a purchase order (PO). So to start, Ariba was a catalog plus a workflow tool, it unlocked the ability to issue the PO to the supplier. Plus the team had designed Ariba so you just got online and it was pretty obvious what the end-user needed to do. At the time, this was pretty revolutionary, the consumerization of enterprise apps was not a thing yet.
Later on, through the network, Ariba gave the supplier different ways that they could choose to receive the PO and made it easier to upload a catalog. But at the beginning, it’s more about the catalog and the workflow, and allowing desktop procurement within a company to happen. Which doesn’t sound revolutionary these days, but was revolutionary in ’96.
Phil: Simple but amazing stuff, Bob. In the language of our Vertical SaaS Knowledge Project, I think about procurement as such a natural control point.
Bob: Procurement is a control point because its workflow cuts across most employees. It has real data gravity because one of these systems becomes the place where the best data exists on where a company is spending money—which is often 50% or more of revenue and costs. And it came along at a time when it was part of the toolset the Chief Procurement Officer brought to bear within a company and CPOs were getting recognized as a key part of the C-suite.
Phil: It also seems like if you land in category X, you can drive penetration in other categories.
Bob: We started with a catalog, and catalog spend is really an item, as opposed to a service. We were better for items than services at the beginning. Something where you can issue a clear PO and get a clear invoice back and you can settle and pay for it.
Could we have expanded to other spend segments? That’s the million-dollar question. If you think about the spend that a company has, a company has its cost of goods sold and its below-the-line costs. A lot of the cost is with having employees like payroll, and an obvious area we talk about in the VSKP. And then there’s all the stuff that’s not payroll. Some of that is not really discretionary spend. Eventually you get down to everything else which is procurable spend.
But each category of spend is different. Consider T&E. It’s a very different spend process. And it’s such a great example of how you have to think vertically to win. We had a T&E module at Ariba and so did Concur. Theoretically we could have succeeded there, but we didn’t make the commitment to the category. You needed to do credit card integration, and have a booking tool, and other things that were very category specific. By not focusing, we sort of ceded that category to Concur. And they ran with it and sold to SAP for more than Ariba!
Fieldglass in temp labor is sort of the same thing. I give SAP a lot of credit for saying we can put all these together into a nice stack. And on top of that, they’ve added a lot functionally for direct. So for their SAP universe, they’ve got a pretty complete solution.
Merchant Side Network Effects
Phil: Using the VSKP framework, how did you start building up density on the buyside?
Bob: It sort of happens naturally. Once you get good at solving the procurement problem for one of those players, then of course a great enterprise sales team goes out to the market with a reference and data points to sell to another enterprise.
In pharma for example, through reference selling, we got eight of the top ten pharma companies in the world on the Ariba platform. Once you have that kind of shared content and suppliers, you’re going to win the next pharma customer. The buyers can see that you understand their business and you can give them a sense that it’s going to work.
It’s also why vertical focus really matters. If there’s enough vertical specific focus at the beginning, you will get enough repeat suppliers.
Phil: A big part of your mandate at Ariba was building out the network. What was your landing wedge with suppliers, how did you get them on the platform?
Bob: The landing wedge was actually pretty basic. We were helping them do business with the largest customers in the world, and convert a chunk of their order-to-cash cycle from paper, phone, and fax, to electronic. Specifically we started with POs and catalogs and expanded from there into invoices, RFQs, contracts, payments, etc.
Phil: How did you build supplier density?
Bob: Supplier density came from two things—having a really good solution for catalog items and having dominance in the buyers who valued automating this spend (high tech, finance, and pharma to begin with). So if you were a supplier selling office supplies, IT equipment, lab supplies, MRO in the US, you were going to keep being asked by your buyers to connect and do business through the Ariba network. We had great density in some supplier types and then an enormously long tail of suppliers across all verticals.
Phil: You mentioned opening up the network to others, that’s a really big decision.
Bob: When I first got there, you could only use our network if you bought Ariba’s procurement software. The analogy at the time was around telco; you can use our network if you buy our handset. That’s not the way most cell networks work. In reality, you can buy whatever handset you want and buy/use the network. So we said let’s sell our network separate from our handset. Let’s let people who are creating procurement POs in SAP or Oracle buy our network separately. That was a controversial decision at the time and took a while for Ariba to accept it. But if you’re going to be a network, you can’t really tell people there’s only one onramp to the network.
It was a big and tough decision. I think this is one of the interesting things about a network is, it doesn’t have any value at the beginning, and then the value grows as the flywheel effect takes effect. So some of it’s just natural. You couldn’t sell the network to people when there was no one on it, right? But then, after your umpteenth customer, we used to be able to go to a customer and say give us your AP file and we’ll show you what percent of your suppliers are already on the network. And that would give them assurance that we were going to be able to make this happen.
So some of it was natural shifting of the value from the handset model to the network. I give Bob Calderoni, who was the CEO, a lot of credit. Bob saw that he had the potential to do an open environment and he was not afraid that if customers had access to our network, they would never buy our procurement tool again. He understood the procurement tool sold on its own. It had its own value, separable from the network, and together they were spectacular. There was a large group of customers, we all realized, that were probably going to buy SAP or Oracle’s procurement solution, but still might want to buy our network. So it was a bold decision.
Phil: Pricing is more than a monetization exercise, it also drives the “game mechanics” of the network. Getting both sides engaged on a common platform is hard, but how did you think about pricing?
Bob: It feels like for the last 20+ years of my life, 30-40% of my time has been spent on pricing. Because pricing is obviously critical to all SaaS software businesses. But it’s even more complicated in multi-sided models.
So let’s take Ariba first. When Ariba started, it sold the procurement tool and gave the network to the customers and suppliers for free. It’s funny, the network ended up being, perhaps, the most valuable part of Ariba, but in the beginning only a few brave souls were willing to work on a product that had no revenue associated with it!
We realized that the network had value that was separable from the tool. Our competitors weren’t selling you a network. So if you bought their software, you would have to build a network or enable suppliers yourself. We realized there was value in the network, we just needed to learn how to sell it and price it appropriately.
Phil: Once you realized there was such value to the network, which side did you charge?
Bob: The first step was to charge buyers. The second decision was to open it up to other buyers. And the third decision, made in parallel to the second decision, was to charge the suppliers.
Phil: How did you know you could charge buyers on a variable basis given your procurement application was not variable, but fixed?
Bob: At first, the charge for suppliers was semi-variable—it was fixed per relationship a supplier had with buyers on the network. This was mainly in the interest of speed but we quickly went to a variable model when we had time to study the value proposition more closely and do market research with the suppliers.
Phil: How did you think of introducing pricing to the suppliers on the system?
Bob: It was a big decision to begin to charge suppliers a small fee to be on the network. That’s controversial even to this day.
Now, when we started charging suppliers, we tried to understand the value suppliers were getting from it. And if you think about what suppliers are getting, it’s the mirror image of the buyer process. We were automating the order to at least invoice—a big chunk of the order-to-cash process, for the supplier. We’re not solving a supplier's complete AR solution, but we’re automating a lot of their orders that would otherwise come by phone or fax.
So we tried to understand how much value suppliers were getting, and what was a reasonable fee to charge. We could also look at substitutes: If suppliers were getting their orders via EDI (electronic data interchange), we’d understand the cost. EDI is pretty antiquated technology, it’s not real-time, there are real issues with it. We dug deep on how much better our invoice process was, how many fewer errors, and how much faster it was vs. EDI, fax, and email. And of course the comparison to [snail] mail was easy.
So we were able to show that if a PO and an invoice all went through Ariba, there were fewer errors—and every time there’s an error, there’s potentially a return or a delayed process. We proved to suppliers that an order through Ariba was much more likely to be a perfect order (PO through invoice that matched and got paid) and much faster than the alternative methods (email, fax, phone, or mail).
Phil: There’s more dimensions to value than just order efficiency, right?
Bob: That’s right. Some of your suppliers on your network are not going to have that many buyers on the network. I used to tell my team that if a supplier has only one buyer on our network, we’re not a network to them, we’re a nuisance! But if a supplier has a hundred buyers, now you’re a network to them. Or you look like one and act like one, and bring that value. So you have to shape the model to the profile of your suppliers.
Phil: Pricing is oftentimes emotional. Was there any blowback?
Bob: Tons of blowback. After all, going from free to non-free is really hard. And it’s easy to criticize the model. It’s less easy to recognize if you’re going to charge suppliers, suppliers become customers. And if they’re customers, you’ve got to think about what value you’re adding for them. In addition to the buyers on the network, we added phone support, we added localization, we improved their interface to the network and added other features to make sure suppliers felt like they were customers. Our competitors would say, “well we don’t charge suppliers.” And I would say, “yeah, that’s true, but you don’t give them anything either”. I used to go do that Dunder Mifflin thing where we call our competitors to see what service we get. It would be nothing—we’d go to the website, there was no place to call, there was no connectivity at all.
And this gets me to the broader point we talked about, which is adding value to the transaction. It’s one thing, for instance, to automate an existing transaction between a buyer and supplier. But can I help a supplier get a new buyer? That’s a different form of value. That’s a marketplace form of value.
Payments and Fintech
Bob: After I left Ariba, some really bright people introduced Ariba Pay. It was a little early though. SAP just bought Taulia, so they’re obviously going deeper into it.
It’s worth framing that we had large buyers, and obviously suppliers of all sizes. But in a lot of transactions, you have a large buyer and a small supplier. You have relatively extended payment terms. Large corporations have been pretty aggressive over the last 20 years with extending their payment terms (40-60 days and they don’t always pay on time). If you have a supplier sitting with an approved invoice on day 10 waiting to get paid on day 45, that’s where supply chain finance and factoring and all sorts of things come in.
While I was at Ariba, I played around with some financing and factoring type stuff—getting suppliers paid earlier. That’s a whole other layer of value. We could add credentialing. We could add a whole bunch of things that I know you know about. And they fit within Tidemark’s VSKP model. The layer cake there, that you and Dave talk about, is similar to the layer cake in these two-sided platforms. Let me look at the transaction and find ways I can add value, and to your point earlier, some transactions you can’t add a lot of value. They’re really simple. But the more deep that transaction is, the more complicated it is, and the more it is between two parties… The more it’s dragged out, and there’s a high cost of capital on one side or the other, the more there are rich opportunities to add value.
Followers of the Ariba Legacy
Phil: You helped to build one of the largest B2B networks at Ariba. Where do you see similar opportunity?
Bob: I loved Ariba, but when I left I told myself that the next thing I was going to do would be less horizontal and more vertical. Since 2010, almost everything I’ve done has been in deeper, more complicated transactions, that the Aribas and the Coupas and the Jaggaers of the world, may touch, but probably not. Because there are other aspects of spend.
I eventually became president of a little company called ServiceChannel, which is now part of Fortive. It was focused on services performed on site. You couldn’t be more opposite of catalog items than that.
And we could go on and on, in legal, in oil field services, in clinical trials, meetings and events, in a lot of these areas where there’s a lot of spend and a really complicated process. So I guess I’ve just had a ton of fun seeing how much value can be added. Yes, they’ll never have a trillion dollars flowing through them, but they might have a couple hundred billion. And the take rate may be really high because the value you can add to that transaction may be enormous.
I have this sort of mental map I use where I think of the y-axis as being the take rate and the x-axis as being the amount of spend that a company can impact. I’ve always sort of put the world’s networks and platforms on that graph. I try to think about where this thing is going to end up, long run, on this map. So you can make a ton of money at a very low take rate with huge volumes flowing through. That’s Visa and MasterCard—they are the rules and the rails but they've got acquirers and issuers taking credit risk. They’re taking their 15-25 basis points, but they’re taking it on trillions.
You can also make a lot of money with a really high take rate on a small TAM (total addressable market). That’s what a lot of marketplaces are. Because they’re, in effect, taking a commission on new buyers and suppliers.
So I look to see if this marketplace solves all this chicken and the egg problems—are they going to find their way up that y-axis and out the x-axis. The upper right is nirvana, where I’m taking a really healthy take rate and I’m taking it on a lot of spend. So I’m always looking at where we can put together a sequence to get there. That’s what I love about your VSKP model—it’s the same idea but we’re thinking about it here with a two-sided equation. Can I build that layer cake of reaching towards the customer, reaching towards the supplier, reaching into payments, reaching into payroll? To me, that’s a little bit like the stacking of the take rate. And the x-axis is just how horizontal can I make it, or how many verticals can I take on.
Phil: I love it, Bob. Well, thank you so much for taking the time. It’s always a pleasure and we’re so happy to
have you as a part of the Tidemark community. Thanks again, Bob.
Bob: The pleasure of working with Tidemark is all mine. I love the way you all think about businesses. And we’ve been lucky to work together over the years on deals.
Share your thoughts
We love the idea of bringing together a community to explore the boundaries of Vertical SaaS and are excited by what we can learn from each other. If you have thoughts or comments or want to get involved, reach out to us at firstname.lastname@example.org. If you would like to keep updated as we publish these essays, sign
The information presented in this post is for illustrative purposes only and is not an offer to sell or the solicitation of an offer to purchase an interest in any private fund managed or sponsored by Tidemark or any of the securities of any company discussed. Companies discussed in these posts may include current Tidemark portfolio companies and/or prior investments made by Tidemark employees while at other investment firms. These companies identified above are not necessarily representative of all Tidemark investments, and no assumption should be made that the investments identified were or will be profitable. The information in this post is not presented with a view to providing investment advice with respect to any security, or making any claim as to the past, current or future performance thereof.
Ariba: Supplier Network
This chapter is part of:
Case Studies and Reference