Tuesday, July 3, 2007
Welcome back to my series on what I consider to be the "key DNA" elements that your product and company must have to be successful in the infrastructure software market. In this blog entry I talk about DNA element #4, which is: try to make sure that your product really plays in a large market that can be penetrated in a cost-effective manner.
[Editor's note: In case you are new to this series on what it takes to be successful in the infrastructure software market, in prior posts I discussed the three prior DNA elements:
|#2 is:||Make sure prior to actually developing the product that you validate that your solution solves a significant point of pain in a relatively frictionless manner.|
|#3 is:||Make sure you build a product that is not a point solution or a feature but a full blown product that can morph into a mini-platform unto itself]|
Size Does Matter
Given that most entrepreneurs at startups and product managers at larger software companies believe in the capitalist system, it is clear that most people in the infrastructure software space want a flagship product to be able to penetrate large and fast-growing markets, as obviously 40% market share of say $1 billion is a lot bigger than say 40% of $20 million.
Given that Venture Capitalists are also capitalists (hence the name ), it should come as no surprise that this question — "Are you addressing a large and fast-growing market?" — is one of the top two to three questions a VC will want answered within the first five minutes of a startup's pitch. Examples of this include:
Sizing the Market
So we get the message: Have your product go after a large market, because you can make more money and it gives you room for error. But how do you really know for your own sake (before you make that three- to five-year commitment of your life) that your idea will equate to a product that truly can play in a large market (and you can make a lot of money)? And once you convince yourself there is a large market out there for your product, how will you be able to clearly articulate to outsiders (e.g., potential investors or, if you are at a larger company, a General Manager whom you are pitching this idea to create a new product team) that a big honking market does in fact exist for what you are proposing?
Unlike talking to 20 to 40 people to get good market validation on whether or not your idea can turn into a "must have" product (see my prior post on this topic), you would need to talk to 100s or 1000s of people to truly determine if a large enough market exists for your product idea. This is probably not going to happen in a reasonable period of time. So how can you validate that your product idea will play in a large market?
First, you can dig up analyst reports from number crunchers like IDC and Gartner Dataquest. If you are three guys in a garage you probably don't have $25,000 to buy a report from IDC, so you probably need to creatively scrounge around and get a few of the relevant numbers that the analyst firm does make public.
So assuming you do find the market sizing data from an IDC that you think may be relevant to your product, my advice is to be highly skeptical of the data and don't use it as the sole justification for the size of market you want to go after. Frankly, most VCs and folks who have been in the industry a while are highly skeptical of analyst market sizing reports. I recall when I was at NetIQ, an industry analyst published a market share report which showed one of our competitors having X% market share of the $Y million Windows management market that we played in. The problem was that the competitor was a public company, and X times Y was greater than their published revenue. Guy Kawasaki lists usage of analyst market share data as one of his top 10 lies an entrepreneur tells a VC when he says: "Every entrepreneur has a few slides about how the market potential for his segment is tens of billions. It doesn't matter if the product is bar mitzvah planning software or 802.11 chip sets. Venture capitalists don't believe this type of forecast because it's the fifth one of this magnitude that they've heard that day. Entrepreneurs would do themselves a favor by simply removing any reference to market size estimates from consulting firms."
Relying on market data from a high-level analyst firm also lulls you into a sense that you are playing in a large market and, oh boy, if you just get one percent you will be rich, etc. The reality is that your initial product will only play in a small subset of the market that has been identified by the analyst, and they are unlikely to have that level of granular market data. I have seen startups pitch that they are part of the $7 billion per year systems management market, but in reality the market is probably only half that (factoring analyst data inflation), and they may only offer, say, a job scheduling product, which is only about 10% of the uber-systems management market, and their product only works on, say, Windows and Linux, while a huge chunk of that particular market is on the mainframe. So $7 billion becomes $175 million real quick if you peel back the numbers, and it will take four years and four versions to be able to fully participate in just that $175 million market.
The second way to do due diligence on your product's potential market size is to analyze your potential attachment rate to the various units being shipped in the future and out there already that your solution attaches to. Infrastructure software tends to attach to a specific hardware and/or software platform - for example, an archiving solution for Microsoft Exchange - so you can figure out how many units have shipped over the last few years and what are the projected number of shipments over the next 1-2 years to calculate the total of units you can possibly attach to. At Centrify we attach to UNIX, Linux and Mac systems, as well as J2EE web servers and other platforms such as Oracle and DB2, so I spent a lot of time trying to get shipment statistics as part of my own due diligence on the market prior to forming Centrify.
But again, don't be lulled into thinking you can realistically attach to all of these units. Say you have a product that is an enterprise software solution for the Mac environment that does backup for the Mac. I believe the currently install base is something like 12 million Mac systems, and Apple is now shipping over 1.5 million units per quarter. So in two years you may think you have a market of 24 million units that you can sell into, and if you can get 25% of that market at $100 per unit that is a $600 million market you will have in 2 years. Not bad. But hold on, what realistic percent of those Macs are going to enterprises, vs. home users or SMBs or education? And what percent of the new shipments are replacing existing systems? And what percent of the Macs are going into geographies that you can realistically sell into (e.g. North America) versus geographies that will be tough to sell into initially (e.g. APAC). And are you being too aggressive in your 25% market share in two years? And can you really hold on to that $100 price point when you factor in resellers/distributors and large accounts wanting 60% discounts, etc. Point is: Ask yourself the tough questions about install base and your ability to attach to it, and if you don't, the VCs will.
The third way to analyze your product's market is to look at what your perceived competitors are doing in the market. If you are the upstart to an established vendor, like Xensource to VMware, you can point to a billion dollar market. If you don't have direct competition, then look at what other vendors are doing in comparable markets and reasonably extract from there. In the mid-90s at NetIQ, when we were really one of the first companies to deliver a performance management solution for the Windows server market, we looked at what vendors such as BMC and HP were doing in the UNIX market as "comparables" because there was no market data for the Windows market, and extrapolated what the Windows market would like in a few years vis-à-vis the UNIX market.
Finally, I think the last way to size a market is to look at major mega-market trends (e.g. rise of SAAS, open source, the growing need to address compliance, etc.) and ask yourself if your product can be caught up in that tornado. In the end, I think most VCs rely on their gut feel that certain mega-trends are happening in the industry that in theory are reshaping the IT landscape, and are looking for companies that are like Dorothy's house in Kansas that will be transported to the land of Oz. Sometimes you need more than one tornado to give yourself and potential investors a warm fuzzy that there is a large enough market out there for your product. When pitching Centrify to VCs, I emphasized how Centrify was at the intersection of three mega-trends: the rise of Linux as a significant player in the server market, the exploding market to address compliance, and Microsoft's continued march in the data center with Active Directory being the key underlying security infrastructure of the Windows platform.
This whole market sizing exercise is an inexact science, but if the numbers don't add up in terms of reasonable attachment rate or what comparable companies (or competitors) are doing in the same or like markets, or if you are not riding multiple tornados, then you may not be going after a large enough market to build a "sustainable, enduring company."
But what if you already launched your product and things are not ramping up? How do you know if the problem is that the market you are in is not a large and fast-growing one, or if you have the wrong sales exec (or CEO or VP of Marketing or team in general), or if your product's feature set is missing the mark? I saw this great blog post entitled "Isolating Causality: Bad Market or Bad Company" by Will Price, which does a superlative job on discussing this topic, so I will leave it you to click and read if you want more information.
But Can You Penetrate the Market?
Finally, so you think you got a great big market, it passes the smell test in looking at size of competitors or comparables, or looking at attachment rates, or riding a bunch of mega-market trends, and the analyst reports are not too outlandish or too small, etc. But that's just part of what you need DNA-wise as it relates to the market. Not only does the market need to be huge/large, but there must be a path to go after it that is realistic and cost-effective.
For example, the Small to Medium Business ("SMB") market for just about everything in infrastructure software is probably huge. But it is very tough to scale because it takes high volume and distribution. So you go Open Source and hope for the viral effect to reduce distribution costs. But Open Source probably only has a 3% conversion rate from free to paid, and will a SMB go for Open Source as opposed to a packaged solution? Likewise, the enterprise market has concentrated and rich customers, but an elongated sales cycle, and the buyer has become resistant to buying from startups. You get the idea.
So the point is: a large market and the go-to-market plan must go hand-in-hand, and often times a VC will agree with you that your product's market is huge, but just does not see a path forward for building a real business in this large market. Sure, sales models can change (going after SMBs versus. enterprises, or switching from direct to indirect sales) but many startups don't have the luxury (or cash) to switch strategy in mid-stream and wait one to two years to see if the new model works, even if it is a great market.
I will blog on go-to-market strategies in the future, but my next posting will be on the next DNA element that I think is important for building a successful business in the infrastructure software market: you must deliver a product that has clear and lasting differentiation from the competition.