How to build a startup: 7 tips from Keith Krach

Entrepreneur Keith Krach recounts how he started Ariba, sharing a treasure trove of tips in part two of this three-part Q&A.

Technology entrepreneur Keith Krach, 62, is on the cusp of what could be a new chapter in his career. In January, the former chairman and CEO of DocuSign, an e-signature software maker, was nominated to be the undersecretary of state for economic growth, energy and the environment by President Donald Trump.

Meanwhile, awaiting confirmation, Krach has been doing some looking back. In part one of this three-part interview, he reflected on his early days at General Motors -- dubbed "Generous Motors" by his father after the automotive giant awarded Krach a scholarship to Purdue University, where he earned an engineering degree, and sent him to Harvard Business School.

Krach also recounted his decision to leave the automotive business for the uncertain territory of Silicon Valley, where he started Rasna Corp., a maker of a new category of software, called mechanical design synthesis. He ultimately sold the startup for a small fortune.

Part two of this three-part Q&A chronicles the hard-driving period between 1995 and 2009, when Krach co-founded the B2B procurement network Ariba. Bought by SAP in 2012 for about $4 billion, Ariba is the world's largest business-to-business trading platform. Here, Krach talks about how to build a startup that wins customers and generates revenue from the get-go.

After the sale of Rasna, you joined the venture capital firm Benchmark Capital as entrepreneur in residence -- on the lookout, I presume, for deals?

Keith Krach, former chairman and CEO of DocuSignKeith Krach

Keith Krach: At Benchmark, we looked at about 100 of what we called mediocre internet deals: Build a website, and they will come. They all said, 'Oh, once we get viewers, we might get advertisers.' It just seemed to me that not everybody was going to be able to do that. I imagined there would be a handful of companies that would make a lot of money that way, but not a huge number of companies. That's what ended up happening.

I had a different idea. I reached out to some of the superstars from the old Rasna team, and I said, 'Look, there's a company in Boston called Open Market. They call themselves a business-to-consumer electronic commerce company. We're B2B guys. Why don't we do B2B e-commerce?'

We also pulled in the second entrepreneur in residence at Benchmark, a guy named Paul Hegarty. He was vice president of engineering at NeXT, so he had worked for [Steve] Jobs. At that time, he was probably the best object-oriented programmer in the Valley. I said, 'Let's see if we can put something together.' We grabbed a couple other guys, so there were seven of us, and we raised some money. We didn't even have a business plan.

When you went to the Rasna guys, you didn't go to the engineers that were the algorithm experts?

Krach: I went to the business experts. We needed people who understood UI and workflow. From Rasna, we got our CFO, vice president of sales and vice president of strategy. And we also found a couple other guys who were thinking of doing something similar at [venture capital firm] Crosspoint Ventures [Partners]. One of them was the vice president of engineering from Home Shopping Network. He was a 26-year-old named Boris Putanec. Boris and Paul formed the nucleus of an engineering dream team that included Craig Federighi, who now runs all software development at Apple.

When we raised money, we brought all the Benchmark partners together and said, 'This thing is worth $16 million. That's the valuation. We need to raise $6 million.' We had no prototype, no business plan. I told them, 'One of the things I learned from you guys is you don't read the business plans anyway, and you don't believe any of the numbers in them. We've got a great team. We know how to execute, position and focus, and there's a big paradigm shift going on in the marketplace. We're going to write the first enterprise application on the internet -- in Java.'

Just based on that alone, you came up with the $16 million valuation?

Krach: We were guessing, but I was confident that we were onto something. I had what I called the 'magnificent seven' factors that I learned at Benchmark going for us. Based on that valuation, we raised $6 million.

How to build a startup based on the 'magnificent seven'

Are the 'magnificent seven factors' some secret formula for how to build a startup?

Krach: You could say that. The first factor is to have a big market. It's just as easy to start a company in a big market as a small market, particularly if there's a paradigm shift. At Rasna, we were in a small market. This was the big one. This was commerce. This was disrupting one of the biggest, oldest industries in the world.

The second of the magnificent seven was positioning and keeping it simple. At Benchmark I talked to a lot of companies that needed CEOs. One founder laid out his company's very complex value proposition, but I liked it. I came back to the Benchmark team all excited and said, 'This looks like a good investment to me.' They asked me what the company did. It took me 10 minutes to explain. They gave me a puzzled look and said, 'What does it do again?' They helped me understand that if you can't position it in a few words, there's probably no market for it.

The positioning out of the gate for Ariba was the 'purchasing agent for the purchasing agent.' That was it. At the time, there was software for finance, ERP and HR. But there was nothing for procurement. At Rasna, we would get frustrated because VPs of engineering would want to buy our software, but the deal would get hung up in purchasing. So, I had an idea: Why not make software for the purchasing guys? It'll never get stuck in procurement.

The third of the magnificent seven was focus, focus, focus. Focus on a defined beachhead, because if you're going to be the category king, you need to start by dominating a market small enough that you can own it, but big enough you can live off of it. Ideally, it's defensible, and it will be a strategic high ground you can use to win adjacent markets.

At Ariba, we narrowed our focus in a number of ways. Instead of going after both sides of B2B e-commerce right out of the gate, we decided to focus on either the buyer or the supplier. It wasn't clear which we should choose. Fortunately, we decided to follow the golden rule -- i.e., the guy with the gold rules. And we focused on the buyers.

Then, we decided to focus on indirect materials instead of direct. Indirect goods are the things you need to run the business that don't go directly into your product -- paper clips, computers and so on. We decided to focus on the indirect side, because companies buy billions of dollars in indirect goods, but they don't have to be made to order. One company's box of paper clips is the same as any other company's. And there was a lot of MRP [materials resource planning] software out there.

Finally, we decided to focus on the requisitioning process, because that was paper-intensive and time-consuming. I had a lot of experience with that at GM [General Motors].

The fourth was execution. You make stuff. You sell stuff. You satisfy the customer. You can have the greatest strategy in the world, but if you can't execute, you're going to fail every time. You can have a mediocre strategy, but if you can out-execute your competition, you can win.

The fifth was the business model. We knew we were not going to be like the 'mediocre internet' companies. We were going to create a real quantitative value proposition and charge a lot of money for it -- the so-called old-fashioned way of doing business.

The sixth was the company with the best people wins. The only thing that the best people need is time. In the business world, time is money.

The seventh was that you have got to be good at raising money. So, we started raising money. That was our pitch. But [Benchmark] had some doubts. I told them, 'You guys have two days [to make up your minds].' Andy Rachleff one of the Benchmark team said, 'No. You guys don't have anything. It's too big of a valuation.'

My answer was, 'You've got two days to make up your mind. Otherwise, we're going over to Kleiner [Perkins], and that would be kind of a shame for your first and second entrepreneur in residence to get funded by another VC firm.

We got the money, and then we decided we'd better do an off-site and figure everything out. We affectionately called it 'the trailer park off-site,' because the space we held it in was pretty low-rent.

How to build a startup: Ask customers what they want

What came out of the 'trailer park off-site?'

Krach: I said to Paul Hegarty, 'Why don't you build the prototype that we can look at. He wisely said, 'No, I'm not going to write a single line of code until we go out and talk to customers.' I said, 'How many should we talk to?' He said, 'Like 50.' I said, 'What questions should we ask?' And he replied, 'You're the CEO; you figure it out.'

So, we identified a diagonal slice of Fortune 500 people to talk with, from CEOs to VPs, to purchasing agents, to individual contributors. We just asked two questions. First, we asked, 'How are you doing it [procurement] now?' They would typically say, 'Oh, man, it's a mess. It's a paper-based process; it takes forever. You never know what you're going to get. I hate the purchasing guy.'

Then, we asked, 'What would be ideal?' 'Oh, it'd be automated,' they would say. 'You could have a walk-up user interface; you could parametrically change the business rules. It'd be integrated with the financials. Maybe someday, it would use this thing called the internet and maybe hook up with the suppliers. You wouldn't need any paper ... So, what are you guys building anyway?'

'That's exactly what we're building,' we said.

Then, we brought 12 of those original 60 [businesspeople interviewed] together. A few weeks later, we had them all present their ideal purchasing solution in front of us. At that time, we had eight or 10 people in the company. We listened, and it was all very similar.

How did you get them to show up?

Krach: First, most of them were in purchasing. Nobody came to see those guys. Second, who doesn't like to share their problems? We said we were building their ideal software, and we wanted them to be part of the foundation. We built up trusted relationships doing that, and it was great.

How to build a startup: Pricing the product

So, before you even wrote a line of code, you're connecting with these customers and listening to all their problems. You're building trust with them, and they become sort of like your beta testers.

Krach: Yes, although we never did betas. We created a corporate account program. Many of those original customers came from that 12-person group.

We said, 'We're only going to pick five companies to work with for this corporate account program. As part of this program you're going to get a discount on the application. You're going to have direct product inputs. We've got to make you successful. What you're going to give us is that you continue to serve on the advisory council. We want you to issue a press release, write an article, gives us quotes.'

A lot of them wanted to do it, so we decided to make room for more.

How did you price it?

Krach: To be honest with you, we weren't sure how to price it at the start. When we finally did do our own internal business plan, the only data on pricing was based on Cisco, because they were writing their own procurement system.

We found a couple other companies who had done this [for themselves], like SGI Sun and Tektronix. I asked them, 'Why are you writing your own?' They said, 'The amount of money you can save is unbelievable. You can channel all this purchasing power, you eliminate maverick buying, and you can negotiate better deals.' So, we asked, 'Why don't you buy it from some other company?' 'Well, none exists,' they said.

So, we had three magnificent factors going for us. First, all the requirements were the same regardless of industry. That told us that it's a horizontal app. Second, companies were doing it themselves because there was such a huge ROI. The third thing was that there was nobody else out there doing it -- yet.

At that point, we knew we really had something. We had been thinking of starting our price at $100,000, but the strong ROI and the fact that companies wanted the solution so much they were building it themselves told us to go higher. We knew a lot of procurement people had authority to purchase up to $250,000, so we set the price at $245,000, and we knew we were still offering a great value.

So, we said to our advisory council, 'Let's meet next quarter, and we'll show you the progress on the product we're developing.' That next quarter, we added 12 more people, so [the customer base] was at 24. And we did it quarter after quarter. It went to 50, 100, 200, 400, 800, 1,600 and 3,200 people. The next thing you know, it was 10,000 people. By then, we called a user conference, and we held it in Miami.

What year was this?

Krach: We started the company in September of 1996. Two or three quarters later, and we raised 6 million in a first round [of public funding]. Then, about a year after we started the company, I thought, 'We need to raise more money.' We had only spent $3 million of the money we had raised earlier, but my experience at Rasna taught me the best time to raise money is when you don't need it.

When it came to valuation, we were in a great position. We had built the product in six months, and we got revenue. We charged our first customers, so we were cash flow positive from our second quarter of existence.

They asked, 'How much do you think you can value [the company]?' I said, half-joking, '$100 million.' This is before big valuations started happening. So, we raised about $18 million in Round B. We had $21 million in cash in the bank, so that meant we only burnt through $3 million in equity. That's equity efficiency. We went public, and our stock tripled or quadrupled the first day. We were $6 billion after two-and-three-quarter years.

Eventually, our valuation got up to $40 billion, because we were doubling revenue quarter over quarter for 12 quarters.

How to build a startup: Revenue recognition

You were cash flow positive the whole time through?

Krach: The whole time. Our CFO and I decided to apply another lesson from Rasna and go as conservative as we could when it came to revenue recognition. We accounted for expenses as early as possible. We didn't amortize any software development or anything like that. And we pushed revenue recognition out as far as we could.

Our CFO said, 'The good thing about the rules for revenue recognition is that, however you choose to do it, you just have to stick with that.' So, we would actually get all our cash upfront, but we would recognize the revenue over two-and-a-half years. We said these would be big implementations, so it would take a while. When it came to financial reporting, we were showing big losses, but we were actually cash flow positive the whole way. The analysts, of course, knew that.

What's the benefit of showing big losses and being cash flow positive?

Krach: Here's the advantage: When 2001 rolled around and the market started to drop, our revenue was still growing. It also gave us tremendous visibility and predictability in the business. The three key things that drive market valuation are cash flow, revenue growth and predictability. That conservative approach meant we had them all, big time -- particularly the growth.

And having been there seven years, you left again -- this time sitting on more money than you had made on Rasna.  What did you do -- come up with 1,000 personal goals, as opposed to the 100 you set when you sold Rasna?

Krach: No. I decided it was time to give back. I was also a little tired. Many of the leadership lessons that I applied at Ariba came from the Sigma Chi fraternity at Purdue, so I wanted to give back to the fraternity. I became the international president, and I brought in all the things I had learned in building companies.

Editor's note: In the third and final part of this biographical Q&A, Krach talks about bringing 'transformational leadership' to DocuSign.

Next Steps

Where modern leadership goes wrong, according to Retired Gen. Stan McChrystal

Asking great questions is how business leaders become great problem-solvers 

Competency is the No. 1 leadership trait needed in times of crisis

Dig Deeper on CIO strategy

Cloud Computing
Mobile Computing
Data Center
and ESG