CEO Spotlight: Cyrus Hadavi, Adexa Software
By Angel Mehta, Managing Director, Sterling-Hoffman Executive Search
Angel Mehta: Let's talk about the early days...Where did the original intellectual property for Adexa's core offering come from?
Cyrus Hadavi: I was at the University of Michigan for about 5 years or so until 1983 when I got my Ph.D. and then started working for Siemens Research Labs in Princeton, New Jersey. Back in 1983, I started to combine ideas from what was known at the time as Artificial Intelligence, and began applying them to real-life manufacturing issues. In other words, how can one create systems that can enable human beings to make the right decisions, work at a very high speed using some of the Artificial Intelligence techniques, and so on. That led us to build a prototype which essentially was a building planning system for very complex environments - namely, semiconductor manufacturing and semi-conductors in general. You have 800 different operations as opposed to 3 or 4 in typical discrete manufacturing - a problem that nobody had been able to solve before, so we actually created the algorithms as well the system for Siemens internal operations. So it was a successful transfer of technology from research labs into an actual working environment....
Angel Mehta: That doesn't happen often....
Cyrus Hadavi: No, it doesn't - so it was very exciting for all of us. Our success basically continued inside Siemens because that same system was upgraded and transferred into other divisions of the company. Anyway, by 1992, I left Siemens and joined a small company (small at the time) called i2...I think was employee # 12. I was there for two years, running implementation and professional services. The company grew very quickly, as you know Angel...
Angel Mehta: So why did you leave? I can understand leaving a company that wasn't doing well, but given how quickly i2 was growing....
Cyrus Hadavi: Back in 1994, I felt strongly that the supply chain problem as a whole is very complex and needed to be looked at differently than i2 was doing. That was one reason. The second reason, which was equally as important, was that I always wanted to start a company....I was afraid I'd become 80 years old and look back and wish I had done it. So I did.
Angel Mehta: Let's talk about the supply chain market in general....it's littered with the corpses of dead supply chain vendors right now. So I'd ask, do you think that when venture firms poured all this money into the different supply chain plays, did they over-estimate the size of the market, or the severity of the problem? If not, what's going on?
Cyrus Hadavi: Well first, I don't agree at all that the market for supply chain is saturated. It depends on how you view the category. Sure, there are a lot of companies in the supply chain space that are very small ... or that there used to be, anyway. But these companies were mostly in supply chain event management - which is a very simple problem. It's a system that can essentially send messages when something goes wrong. It's just about tracking and signaling what they need; any database can do that, and that's why you've got such a proliferation of these companies. They all died (or are dying) because there was no real value embedded in it. What we do is essentially modeling the manufacturing environment and the supply chain.
Angel Mehta: What exactly do you mean by modeling?
Cyrus Hadavi: Think of it this way: human beings are "intelligent" because they have models of the world inside their head. So if I want to go to the airport, I have a model of: I can get there by cab, or I can get there by walking - whatever it is. And it's going to take me 2 hours to get there or 1 hour to get there, depending on which method I use to travel. Based on that model of how I plan my trips, I can plan the future. That's what I mean by "intelligence" - the ability to model and, based on that model, to extrapolate and project the future. That's the difference between what we do and what the supply chain event management vendors do - we give the ability to model the environment and extrapolate a view of the future that allows for decision making. Only a handful of companies even claim to do this. I2, obviously, and Adexa. But the kind of domain knowledge that Adexa has gathered....from over 100 customers....it's not easily available.
The problem with i2 is, they tried to address a very new and complex problem with very old technology. They made a lot of commitments that they could not keep...they over-promised, and tried to deal with it by acquiring lots of different smaller point offerings. Bringing those solutions together was just extremely hard, and it was too little too late.
Angel Mehta: So Adexa is better off for having grown organically?
Cyrus Hadavi: Actually, yes. The solution is a unified data model, which really leads to a much lower total cost of ownership. So back to your question of the market size, with the supply chain problem as a whole, it's not going to go away...It's very fundamental to every manufacturing company, and if anything, it's becoming more complex - not less complex. The current estimates are that 50% of the world's manufacturing is going to be in China. In North America, we're going to have 9 out of 10 people being in the service industry. So these companies in North America...the fabless companies in semiconductors - like the Cisco's of the world - they're going to be relying on a very complex supply chain...Their manufacturing gets done 17 time zones away! They need to have visibility, they need to have planning, they need to move quickly and react to changes in the marketplace. All very fundamental stuff to any company. So the supply chain problem as a whole, if anything, is getting worse - it's not a threat that comes and goes. It does require very deep domain knowledge to solve, which we are fortunate to have at Adexa.
Angel Mehta: On the Adexa website, there's a link that speaks directly to I2 customers....
Cyrus Hadavi: Right.
Angel Mehta: I think it's the I2 exchange program. When and how did that come about? What inspired it as a tactic?
Cyrus Hadavi: It's pretty simple. There were a lot of "rumours" about I2's customer base not being completely happy with what they had received from I2. There was Nike, which was very public. There was K-Mart, and a few others. Then, recently, Nucleus Research - which is an independent research company - tried to identify the ROI of an i2 implementation. So they went to i2's "referenceable" customer base, and independently asked them about results. Fifty-five per cent of them came back and said, "We did not achieve the targeted ROI." So we compared that result with our own customer base, and felt that that this was a huge opportunity for Adexa to go to I2's customer base and show them what can be done, and what should be done. That's really all there is to it.
Angel Mehta: How does the recent consolidation - or at least, intended consolidation - by the larger ERP players affect Adexa? One entrepreneur told me privately that he's upset about there being fewer buyers in the market that he will have a chance to sell out to...does that ever enter your mind? If any of the deals does go through, would it make, say, Peoplesoft or JD Edwards more or less competitive?
Cyrus Hadavi: It doesn't matter. The bottom line is, if you look at the landscape of supply chain planning and the companies that have the domain expertise to address it...there are just very few players out there. Peoplesoft and JD Edwards do not have a comparable offering. If you look at SAP and Oracle...they've declared war on supply chain vendors just like they have declared war on Siebel with regards to customer relationship management. But SAP has been talking about this for the past 5 years, and still, their solutions are far from where they need to be. Oracle has sold to a few small fabless companies....but ultimately, we are in a very unique position from a technology perspective.
Angel Mehta: What is Adexa doing differently now than you were during the bubble?
Cyrus Hadavi: During 1999 and into 2000, our customers used to buy everything at the same time. We were focused on end-to-end solutions -- from business planning to shop floor sequencing -- and enabling our customers to collaborate with their customers and multiple levels of suppliers. Now the question for customers is "where is my immediate pain and what pill (application) can I take to give me immediate relief and ROI at the lowest cost?" This is excellent for Adexa. We have scaleable applications to deal with specific pain-points, and they seamlessly integrate with our other applications when customers want to focus on their other issues. We have a lot more interaction with our customers today, in terms of meeting their specific short term objectives, and the relationships we build in the process serve us well in the long term.
The other thing is, we had a lot of people who had joined the company during the bubble economy...they just left school, and were highly compensated. They were given promotions very quickly...they were given a lot of stock options that obviously, at that time, translated into a lot of money. And when the bubble burst, all of that essentially went away. Educating our employees was a key challenge. Explaining to people that the bubble was just a dream - the reality is what we see today. We had to explain that to build a company, you have to make it through those tough periods - you have to be able to withstand the downtimes. It took us about 2 years to be able to manage people's expectations - some people accepted this; some people didn't. The good news is that people who are still here, they understand what it takes to build a great company. They're not here for short-term gain.
Angel Mehta: Let's talk about the market in broader context...many people seem to concur that the enterprise software market is maturing - that the bigger companies really don't need anything beyond what they've already purchased. I want to ask you about that from a supply chain application context...Is that true? In other words, I understand that there are companies that are not happy with the applications that they've purchased, but does that make up the totality of the market, or is there an entire market out there that hasn't even begun to explore automating the supply chain?
Cyrus Hadavi: First of all, people may have spent a lot of money on buying supply chain management or other types of enterprise software. But they have not really uncovered the value of what it has to offer. Most companies have bought supply chain planning for the internal operations of the enterprise. But the complete solution - a solution that can take demand from customers in terms of their sales planning, compare the internal constraints and then take it even further to the supplier base to see if they can deliver, and based on that cycle back to the customer and make commitments with regards to capacity....that is still a dream for most companies.
Another area where we see a bright future is to expand into the concept of what we call "Self-Correcting Supply Chains." Typically, you have hundreds - if not thousands - of suppliers. Typically, you have hundreds - if not thousands - of customers. You have thousands of resources, etc., so that is a lot of data....a lot of signals coming at you. The real question is: How do I detect trends? How do I detect exceptions? For example, let's say a supplier commits that they deliver in under 10 days, every time. You examine the supplier over a 3-month period, and you realize that 95% of the time, they are delivering in 8 days. Well if a supplier is really delivering to me in 8 days, 95% of the time, I am in a position to adjust my supply chain model. I can now commit faster to my customers because I know I can get the stuff inside my enterprise 20% earlier. Not to mention, my inventory is lowered because instead of 10 days, they deliver in 8.
Better customer service at lower inventory cost is the simple result of monitoring a trend. That's what I call a self-correcting supply chain - where a company can collect the appropriate data, watch for trends, and when a threshold is crossed, use that information and tell the relevant constituents about it. It's a huge segment in and of itself.
There are thousands of pieces like that. Another untapped market is the smaller and medium-sized segment. When you look at countries like China, India, or Russia....or down the road - 10 years from now - the Middle East, despite the current political environmental situation, they have a lot of manufacturing, and we haven't even touched any of those places.
Angel Mehta: One of the things that we try to get every CEO to do with the Sterling report is share 3 key mistakes that they could think of that they've made or the company has made over its history...Can you comment on that in your experiences at Adexa?
Cyrus Hadavi: Sure. One mistake was that I should have raised money a lot sooner. I boot-strapped Adexa for the first 3 years. We were profitable, so I didn't see the need for financing at all. But not raising enough money caused us some headaches down the line because Adexa was not as well known as it could have been and didn't generate as much marketing hype as it should have. So that would be my Number One mistake. The second mistake has to do with not raising enough money when we finally did go out for financing. I was very thrifty about how much money was needed...in retrospect, that may have been a good thing - but early on, I think we could have expanded a lot faster.
Angel Mehta: You talked about the opportunity overseas...how far as Adexa pursued the international market to date?
Cyrus Hadavi: We are very much diversified that way. About 40% of our revenues comes from Asia. We are very strong in Japan...We have pretty much all the blue-chip electronics companies in Japan. We have 25 customers in Taiwan...a number of installs in Korea, an office in Singapore and we have just opened a new office in Shanghai. The international exposure has really helped us afloat in a tough North American market.
Angel Mehta: Interesting. How did it come to be that such a large percentage of the revenues have been from international?
Cyrus Hadavi: It was very simple. I went on a road trip with a partner to Korea and Taiwan. I had a big HP work-station at the time, which I put it in a suitcase with some home-made padding....and I just took the suitcase, went around, and gave demos. We managed to get 1 customer in Taiwan - the 2nd-largest foundry in the world (UMC) as our 1st customer. In Asia, it's an interesting environment. They are very detail-oriented, not to mention consensus-oriented. That want to see lots of substance...they take their time - sometimes up to 6 months of very detailed demos. They're not as open as we are in North America to sales and marketing hype. But once they buy into an offering and the technology works, word of mouth travels really fast. The way it works in Asia is, if you have 1 bad customer, you might as well just leave the island. But with very little marketing at all in Japan, our average sales price was $1 million in software - it still is.
Angel Mehta: Did the high cost of sale pose a problem for you?
Cyrus Hadavi: When we expanded, it was originally through our partners in Japan. In Taiwan, we had 1 or 2 people who went back and forth to make the sales there. But as soon as we had a couple of happy customers, we decided to make investments in those areas and watched it very carefully in terms of how we are doing, how many people we are adding, and so on. I think it was a good strategy. We did exactly the same thing in Singapore where our 1st customer was Chartered Semiconductors and Chartered is the 3rd-largest foundry in the world. Our first customer in Korea was Samsung, Plastics Division. The approach we've had is, first get a customer, and then make additional investments and grow. In China, given the potential that the country has to offer, we had 1 person on the ground for the past 2 years doing business development and building Adexa's reputation and mind share. Just 2 months ago, after 2 years, we in fact went ahead and spent some money opening an office and making the investment, etc. But we have taken a very careful approach in how we approach these markets because it can be very expensive. In Europe, unfortunately, we have not been as successful. I think maybe going back to the mistakes that we made is that we treated Europe as a kind of United States of Europe, and we felt that by putting our people in 1 or 2 locations, we could go after every region in Europe. That was a big mistake because the French, for example, they want their own vertical integration team. Germany wanted the same. So it was a very expensive learning experience for us in Europe, even though at some point in time, Europe was about 20% of our revenue. But for the past 4 years, it has been a huge drain on our expenses. Recently, we decided to cut back.
Angel Mehta: Would you start an enterprise software company in today's environment?
Cyrus Hadavi: Absolutely. There is an incredible amount of opportunity right now to add value. If you look back at the early 1990's, the world went through pretty much the same kind of setback as far as enterprise software is concerned. Up to about 1993/1994, there wasn't much activity. And then, all of a sudden, the supply chain market took off.
The early 1980's were a downtime and then in the mid-1980's, it started to take off again with enterprise software. Before that, there was this concept of, "Don't concentrate on a machine; concentrate on the factory." That caught on for a while, and now we're in the next phase. For the past 70 years, we've gone through a focus on the workers, to a single machine, to a single factory, to the complete enterprise. But the supply chain is just a small piece - now we're talking about what I call the trade value network - which is a huge opportunity. This problem of B2B - Business to Business planning - it's a huge, huge problem that most companies haven't even thought about. They talk about transactions, or they talk about exchanges and sending messages back and forth. But the problem of modeling connected companies - you have finances, you have banking, you have insurance...transportation...and your suppliers!! All of these need to be connected at the modeling level - not just as at this transaction level. Maybe I get so excited because of my domain knowledge here, but it represents a huge opportunity. People always have a tendency to think that the current reality is going to last forever. But that's never true.
Cyrus Hadavi is Chief Executive Officer of Adexa. To send feedback to Cyrus, email: email@example.com
Angel Mehta is Managing Director at Sterling-Hoffman, a retained executive search firm focused on VP Sales, VP Marketing, and CEO searches for enterprise software companies. He can be reached for feedback at: firstname.lastname@example.org