Lookery.com

Image via Wikipedia

Much is written about both whether or not the VC model is broken and how to evolve startup business plans in the face of market changes. Today, it hit me just how inextricably linked the two issues are and how my own tactical process needs to catch up to market realities.

I spent from 1992 to 2006 presuming VC-backed startups were the business I was in, and I worked towards understanding that system. After all that time, two things happened to radically change my outlook.

1. I finally figured out that one should only raise VC if one is already rich,
2. I also figured out that being boring and late has better risk|reward characteristics than being sexy and early, and
3. Cloud computing arrived, making VC deal terms economic only as growth capital for Internet startups, leaving the early-stage field clear to angels (and bootstrapping).

Taking any number of lessons from Mashery’s good work, Lookery is a cloud-hosted SaaS vendor that uses an API to provide deep benefits to its customers and suppliers. Like Mashery in early 2007, we’re actively sorting our which customers and which decision makers love us and which look at us crosseyed (or don’t look our way at all). We have numerous data points in each category and the right kinds of patterns are emerging.

The problem is 15 years of old work vs. 3 years of new work. I haven’t finished retraining myself not to presume VC. I find myself mentally parcelling out multimillion dollar budgets that don’t exist. More importantly, calculating low-capital SWOT is not truly intuitive, particularly when analyzing VC-backed companies in Lookery’s market sector.

The VC-backed companies in our sector (principally Blue Kai and  Exelate) are doing a great job getting and giving data distribution via cookie exchange without the benefit or overhead of a centralized profile hosting system.  Cookie exchange works well for many user-targeting applications, but there are a few key tasks that aren’t covered including:

• Efficient combination of data from multiple sources;
• Forcing and enforcing the anonymization of targeting data without depending on good behavior by publishers and/or ad networks; and
• …

Lookery exactly runs that exact scaled profile hosting system, and it changes the equation — but how in a SWOT context? We’re angel-funded and intend to remain that way until we’ve completely nailed the revenue model (see #3 above). Relative to the other sector participants, our near-term enterprise value calculations and related tactics are different. My erroneous, knee-jerk reaction is to compete directly with them but that makes no financial sense. They have an order of magnitude more resources (from their VCs) and a lot more pressure to scale revenues quickly without much regard for expense (also from their VCs). We certainly grow revenues every month but breakeven in Q4 is a much higher priority than absolute scale right now.

The punchline on SWOT for Lookery in 2009 is to build on the unique strengths of our system putting priority on relationship depth and interconnectedness. We want to be our customers’ profile hosting and delivery system — and the one they want their partners to use. That means of our customers require a little more care and feeding, plus we have to be careful to disclaim all rights to their profile data. It’s business that the heavily funded startups can’t quite slow down enough to satisfy, gives them a good reason to do business with us, but is healthy enough to drive us to scale next year.

Related articles by Zemanta

• Who uses cloud computing? Startups do, VCs don’t (venturebeat.com)
• 5 Milestones to Reach Before Raising Venture Capital (centernetworks.com)
• Without APIs, there is no cloud computing (news.cnet.com)

“It’s important to remember the board’s primary purpose: to hire (or fire) the CEO.”

-

Andrew Payne: How to Run a Startup Board Meeting

Rafer sez:
I love Lookery’s angels.

Building Voldemort read-only stores with Hadoop: At Lookery, we have been working very hard to transform most of our data processing tasks into batch-oriented workflows in order to deal with growth. For example, we were already using Hadoop to compute our index and data files for our largest database, but the process of serving that information took place over too many network hops (load balancers, reverse proxies and Amazon S3). Therefore, as soon as I learned that Project Voldemort supported offline building of distributed stores, I decided to try it and we’re now running it in production. Just visit the guest blog post over that Project Voldemort blog for full details.

Image by akakumo via Flickr

We’ve been out selling Demographic Targeting to ad networks for five months, and the first stage of our Post-Facebook era is going fine. We have happy customers, stable infrastructure, etc., so now we know what our operations really cost. Thanks to @sawickipedia, we priced ourselves correctly for ad network sales, but that’s only a few hundred customers. Now that Lookery’s per-function IT costs and margins are clear, we can work on additional pricing plans with different value tradeoffs to greatly expand our available market.

That magic is that we were able to optimize our serving infrastructure after deployment. Over the course of the last 6 months, we’ve gone much further into Hadoop and Project Voldemort. That means we started with the wrong server count, big-small box ratio, et al. So we just shut them off and turned on the exact number and kind of server instances that we need when we need them. There’s another million bucks we never wasted.