tecosystems

Everyone’s Wants a Piece of the New Guy: The Startup Q&A

Share via Twitter Share via Facebook Share via Linkedin Share via Reddit

Joyent’s David Young, a notable Sun customer and frequent beneficiary of pointers from Jonathan Schwartz’ blog, doesn’t get Sun’s new program to offer discounted pricing to startup businesses. He lists several objections to the idea, some of which I agree with and some of which I don’t. The following, IMO, is the most important bit:

Paul Graham says start-ups can get going on $10,000. How? They aren’t buying Sun. (Not that Sun is expensive. But no one knows that, or believes it. Joyent knows otherwise.) These start-ups are on a 1U server they bought on woot.com with a copy of Ubuntu installed. But let’s face it: if you’re running your start up on Linux and some white box brand, you don’t view infrastructure as a competitive advantage. It is a cost. These companies view their software as the advantage, and they can’t buy that from Sun.

While David’s comments are directed at Sun, their importance transcends any single vendor. This is, in fact, a problem that a host of firms large and small are struggling with. We get asked the question all the time: how do I get startups to build on my stuff? While there’s no easy answer here (as if there ever is), I thought a quick Q&A might flush out some of the issues.

Q: You mention a host of vendors: can you be more specific? Which vendors do you believe are concerned with the problem of selling to startups?
A: Most purveyors of general purpose software are or should be concerned with getting startups using their products. Look no further than SAP’s recent ad campaign, which features a company that “definitely needs to go global – right after it goes national.” Or the success that MySQL has had selling into markets that DB2 and Oracle would not traditionally be interested in. Startups are a compelling market for the vast majority of software vendors I speak with.

Q: Do you feel that the value placed on the market is justified?
A: Well, it depends. In strict finanical terms, no – or at least, probably not. David’s not wrong when he says that startups are not keen on paying for service. To borrow MySQL’s way of phrasing it, they’re typically willing to trade time in order to save money. This is particularly true of bootstrapping firms operating off of shoestring budgets, but it’s also the case for a great many VC backed entities. So the simplistic to the question is no, the startup market is likely overvalued.

But that answer is, as far as I’m concerned, overly simplistic. Google being as anomalous a firm as Microsoft, they’re perhaps not the best case to extrapolate from. So let’s instead look at YouTube. A year ago YouTube was nothing more than a startup absolutely smoking through cash. Today, they’re a $1.6 billion acquisition commanding more and more attention from media outlets seeking to use them as a vehicle to users. They’ve graduated, in other words, from a point where they’d be willing to trade time for money. It seems far more likely that YouTube and its parent Google would be willing to trade money for time. That’s great news if you’re the technology they built on top of, sad news if they’re not. While migrations away from the original platform do happen – Google reportedly tried to cut away from MySQL for its AdSense offering – they’re infrequent and problematic (the Google migration failed). The easiest way to turn a startup into revenue, then, is to get in at the ground floor.

Q: Isn’t this a bit like the lottery, then? For every YouTube, there are a hundred firms that fail. What kind of odds should vendors expect?
A: The answer is yes, it’s very much like the lottery. The odds of any given startup hitting it big are minimal. As one person put it to me this week, “so vendors should be looking at themselves much like VC firms do.” The key here is betting widely and betting intelligently. The good news is that in the software space, there are ways to make the betting easy. Open source, for example, is a terrific strategy for lowering barriers to adoption. It allows you to potentially leverage even those customers that are not paying you, either by the potential of future returns or with soft benefits such as bug reports, forum participation, etc.

Also, just because a startup fails does not mean that efforts to recruit them to your platform do. Jon Shea, in a comment on David’s original entry, explains it thusly:

And imagine this: You’re fresh out of college. You join a startup as a “SysAdmin”, even though your only experience is dorm room experimentation. Your company, 53People.net, goes with Startup Essentials from Sun, because it ties Dell in price, and you’ve read awesome things about Sun on Joyeur. Two years later, 53People.net’s second round of funding dries up, and the company folds. Now your looking for a regular SysAdmin job. And at that new job, you’re going to be a Sun evangelist. Sun could end up with a little fan club sprinkled through Fortune 500 tech departments.

We’ve had experience in this directly, as we often seek to leverage our connections with large vendors to the benefit of promising startups. Some succeed, most fail, but they can all add value.

Q: What should companies looking to court startups do? How should they approach the challenge?
A: There are lots of things firms can do; it’s all dependent on the product and market. But a couple of general suggestions:

  1. Give startups what they want, not what you think they should have:
    LAMP is the dominant stack within the startup communities, so vendors need to accept that. You can compete with an individual element or two, but if you’re trying to push a completely different stack it’s going to be a Sisyphean task. Given them what they want, not what you think they need.
  2. Make their life (significantly) easier:
    As I’ve told a few audiences recently, Amazon’s EC2 & S3 offerings are perhaps the most interesting new offerings I’ve seen in a long time. Why? Because they make (or could, if the tools were better) deployment and scaling a push button activity. The simplest way to think of them is as the hardware equivalent of Software as a Service (SaaS): EC2/S3 are, essentially, Hardware-as-a-Service. This can make a startup’s life significantly less complicated – just ask SmugMug.
  3. Let them trade time for money:
    If you’re a commercial provider competing for the attention of a startup, licensing costs are going to be tough to swallow. Unless the product area is relatively specialized, there are likely to be free, open source alternatives available that you will be forced to compete with. Charging for the right to use the software (RTU) will be a non-starter in most markets. There’s a reason the Oracle FAQ runs MySQL.
  4. Know how to separate the wheat from the chaff:
    Chaff being, in this case, very small startups with minimal purchase requirements. RedMonk, as an example, qualifies as chaff: I signed up for Sun’s Startup Essentials (we just made the cut, as I applied a month before we turned four) so that I could purchase a box or two at a discounted rate if I need them. No more than that. Having a salesperson call me for that is a waste of everyone’s time. Courting the Joyent guys, on the other hand, is probably a good idea seeing as they (want to) buy a lot more hardware than we do. The point is that all startups aren’t created equal, and you shouldn’t treat them that way. Court them all, but differentiate wherever possible.
  5. Speaking of salespeople…:
    Jonathan has argued that it will be important for his firm to be able to market and sell to customers they never touch directly. I happen to agree. I’ve bought a couple of Dell’s and probably half a dozen Thinkpads over the years, and have yet to talk to their respective salesforces. But it goes beyond questions of scale: salespeople are the bane of many an enterprise IT manager’s existence. Unleashing them on time-strapped startups is not likely to win you many friends. Be prepared when they come to you, but limit the outbound side lest you appear overeager.

Lots more we could talk about here, but those are where I’d start.

Q: How would you go about courting startups, if you were in a vendor’s shoes?
A: Well, the first step is to have – as discussed – an offering that’s attractive to them. Trying to pitch them on the uber-server package with all the bells and whistles probably isn’t going to get their attention; free software, free hardware, free hosting, and free services are likely to.

So primarily, my channel of choice would be other startups. Word of mouth amongst smaller, technology oriented startups is hyper-efficient in the new world of blogging, wikis and forums. As the guy responsible for RedMonk’s purchasing, for example, I rely primarily on the experiences of my peers. This presumes, of course, that you have startups that you can work with. In the absence of such candidates, you might look around for folks in need, or if you’re a customer of ours just give us a call – we talk to startups all the time that could use some gift hardware or software.

I’d also frequent and support conferences like those thrown by David Berlind and Doug Gold – the Mashup Camp and Startup Camp unconferences. While these are my favorite conferences because I get to walk around in jeans, my Sox hat and flip-flops (kidding), they’re also not bad places to meet entrepreneurs with the goal of becoming the next Google. They’re essentially The Place to be to meet and mingle with that crowd, which I’m making an exception to my January travel moratorium to attend Mashup Camp in Boston.

Q: You mentioned Amazon’s EC2/S3? Do you think those are promising approaches for startups?
A: Yes. They’re not perfect either in pricing or technology, but they have serious advantages over dedicated, self, or colocation type hosting. Not just in terms of ease of use, but with respect to handling the challenges of scaling, cost containment and more. While I expect to see one or more of the eBay/Google/MSN/Yahoo/etc parties to develop competitive services within the next 12-18 months, I think at this point EC2 & S3 must be viewed by IT vendors both as threats and opportunities. Threats, because startups that base their businesses on this technology could be hard to win back, but opportunities because ultimately HaaS needs software to run on top of it.

Q: Anything else to add?
A: Lots, but it’s a Friday. If you have questions you’d like answered, drop them in as a comment or shoot them over to me and I’ll try and get to them in a follow up.

Disclaimer

: IBM, MySQL, and Sun are RedMonk clients, as is Oracle via the Sleepycat acquisition. Also, Joyent provides me personally with a comped hosting account, while Amazon, Canonical, Google, SAP are not (Amazon’s subsidiary A9, however, is a client).

6 comments

  1. I think the value proposition for a startup might show up after they’ve been operational for 6 to 12 months and they’ve got their software model getting locked in place and are beginning to get their cash flow to work better (well, assuming we’re talking about net positive cash flow operations that I think are the way to avoid bubbling and not giganto-YouTube type deals), at which point in order to sustain growth, infrastructure scaling gains more thought-time in terms of technical problems.

    LAMP though… I know the LAMP stack, explicitly LAMP (Linux, Apache, MySQL, Perl/PHP/Python) was dominant a two years, even a year ago, but isn’t it becoming more like, O/LWDL (Open Source/ Linux, Web Server, Database, Programming Language)? I don’t see too many people talking about Apache or MySQL exclusively in their articles (I could just be reading the wrong ones), and I know PHP is still the bigguns in terms of Languages people are using, but I’m starting to think it’ll only be a few months or maybe a year before somebody lights the world on fire with an app built on Haskell or OCaml or something like that.

    Ehh.. whatever, pardon my insane 3am ramblings. Man, O/LWDL is a *terrible* acronym.

  2. You’re fixated on LAMP – that seems like a radical oversimplification. I see a ton of Windows, and the “L” in LAMP leaving Fedora behind, and even some folks (Joyent’s a good example) picking up Solaris. Postgres is picking up, Cloudscape too.

  3. Danno: on the LAMP front, it usually still is those parts in varying combinations. the point is not that LAMP is the only stack i see, but it is the most dominant. there are a variety of choices on each tier, and different startups make different choices based on their needs, resources, and so on. but the numbers are fairly clear that the individual components of LAMP are still the most popular in their respective categories.

    Francis: don’t mean to be fixated, as i spend a lot of time looking at non-LAMP technologies like Postgres, Ruby, (Open)Solaris and so on. there are a great many important technologies besides the LAMP stack – Windows certainly being one of them.

    but as discussed above, LAMP is still the most common combination that i see within the startup ecosystem. that’s in speaking with them directly, learning more about them at conferences, or speaking with the VC’s that fund them.

  4. […] Interesting factoid I picked up here.. which pointed me here and which pointed to the proof point here. (didn’t even take 6 degrees of separation) Apparently the Oracle FAQ site orafaq.com actually runs on MySQL – not Oracle. That seemed very odd at first, but then it makes sense in the context of the cost of an Oracle commercial license. I’m surprised Oracle has not given them a license “at a large discount”…   […]

  5. […] A month ago today I spent a fair amount of time discussing the business of selling to startups, so it was with interest that I read today’s TechStars news. You can get the news first hand via Alex, Brad or David’s blogs, but essentially the program is this: would-be startups apply, and 20-30 winners are rewarded with $15K in seed money, educational panels and the like, networking opportunities galore, and finally an opportunity to pitch their idea to VC’s. […]

  6. […] The simple part is something I’ve written about before, and talked to Sun about, actually. Here’s what I said last month: Know how to separate the wheat from the chaff: Chaff being, in this case, very small startups with minimal purchase requirements. RedMonk, as an example, qualifies as chaff: I signed up for Sun’s Startup Essentials (we just made the cut, as I applied a month before we turned four) so that I could purchase a box or two at a discounted rate if I need them. No more than that. Having a salesperson call me for that is a waste of everyone’s time. Courting the Joyent guys, on the other hand, is probably a good idea seeing as they (want to) buy a lot more hardware than we do. The point is that all startups aren’t created equal, and you shouldn’t treat them that way. Court them all, but differentiate wherever possible. […]

Leave a Reply to Michael Dolan Dot Com: Linux, Law, Open Source » OraFAQ.com runs on MySQL?? Cancel reply

Your email address will not be published. Required fields are marked *