Thursday, March 27, 2014

Why Scalabilty is Overrated, and the Postmortem Of a Startup

I've spent much time looking back at the 2000 era Dot-Com roller coaster I was part of, know as Amazing Media. The idea of user controlled ads, as we know thanks to Google AdWords, is a bazillion dollar idea. Yet, even with millions of dollars in funding, and a top notch team (if I do say so myself), we couldn't get the idea to fly. Why?

I think Seth Godin's recent blog post, Not even one note, hit the nail on the head:

Starting at the age of nine, I played the clarinet for eight years.

Actually, that's not true. I took clarinet lessons for eight years when I was a kid, but I'm not sure I ever actually played it.

Eventually, I heard a symphony orchestra member play a clarinet solo. It began with a sustained middle C, and I am 100% certain that never once did I play a note that sounded even close to the way his sounded.

And yet...

And yet the lessons I was given were all about fingerings and songs and techniques. They were about playing higher or lower or longer notes, or playing more complex rhythms. At no point did someone sit me down and say, "wait, none of this matters if you can't play a single note that actually sounds good."

In other words, we had an arguably excellent idea, but, we never stopped to make sure it actually worked on a small scale. We never got that one note to sound right.

The talk was always about making the idea scalable. How could we have the software do more for more customers with more sales people selling it. And to some degree, that's fine. But, before we talked about more, before we wrote a single line of code, we should have made sure the idea could actually work. The premise, in the case of Amazing Media, was that any old business could benefit from designing and running an Internet ad campaign.

If we had done this manually; just gotten 5 of our favorite businesses and hand made banners and run them on the web, we'd quickly realize that it simply didn't work. The problem we would have discovered was that for an arbitrary business the ROI wasn't there. Sure, we could show Bob's Flower Shop ad 100,000 times. But, digging deeper, we'd see that were probably showing it mainly on an Auto Repair Forum frequented by folks on the West Coast, when Bob's was in DC. Would it be any shock that Bob's phone didn't ring off the hook? (Over time, we added various targeting features to the system to correct this, but they never quite solved the ROI problem.)

What Google AdWords got right (among other things) is the whole question of relevancy: I'm searching for "flowers in DC" so showing me Bob's Flower Shop is a good thing.

Had we tested the model on a small scale, we would have seen this flaw. We were smart folks, so I'm convinced that we could have figured out a solution. And then all that investment money would have gone to building something that actually works. Instead, we spent time and energy building something that looked good, but never actually delivered on its promise.

I'd like to think that I've taken this lesson to heart. And you can, too. Before you go off and build that million dollar idea, or add that million dollar feature, ask yourself: how can I test this out on a small scale? How can I see if the idea actually works? You probably won't need anything more than a spreadsheet or a notebook to do this. But what you'll learn should shape your plans immeasurably.

NOTE: Please don't take the above analysis of Amazing Media too seriously. I'm sure I've forgotten important details over the years, and I'm sure there were behind the scenes goings on I knew nothing about. Still, I think the take away above is at least one critical reason the start-up never really started up.

No comments:

Post a Comment