Programmatic Advertising: Case Study in Customer Dev.

home_3_2I haven’t talked much about my failures at Deconstruct Media, but I read an article today that made me want to. For many, many entrepreneurs, selling a company is a great time to do a victory lap. I know I did, albeit a maybe more somber victory lap than some. But let’s be clear: When you sell a company pre-revenue, it is one part “YOU ARE SUPER AMAZING” and one part “HOLY SHIT, WE ARE GOING OVER THE WATERFALL”. And when I say that, I mean that someone on the outside looked at you and deemed you amazing enough to be worthy of acquisition and someone on the inside (me) said, “this is not going to work, we need to take this offer”. Selling your company has an element of wimping out to it and that is what I did. I got scared and I took the offer.

I still think I made the right decision. I want to talk for a minute about why it wasn’t going to work, but to get to that, we need to talk about what we did.

Deconstruct Media was a pioneer in what came to be called the Programmatic Guaranteed market. When people looked at our product, they compared us to isocket and buysellads, but our idea was a bit more complicated. We recognized that this was a two-sided market and if the plan was to tax publishers, it would be ugly, but we needed to make one side of the market easy. As Mike Yavondite, who had built and sold a similar business (Quigo) to Aol, told me when he politely passed on my round: “Two-sided marketplaces are hard. Realistically, if you are going to do that, you need to have a plan that makes one side of the market totally frictionless or else it will just be too hard.” This comment made a lot of sense to me, I have repeated it to many entrepreneurs, and it haunted me through the process.

So here was our plan: Sign up publishers to sell inventory on our platform by offering them a simple self-service ad platform that plugged and played with their ad server and could feed our infrastructure standardized data about their inventory, then combine all that inventory into a single forward market exchange. This is consistent with a lot of stories about building network effect businesses: You need to deliver value to the user before others join the network, our value was this tool.

Of course, we thought this was a business that had a network effect in that if we built the demand side, the supply side would want to participate. And vice versa. All we needed was a way to onboard demand and we would have it. We made some progress on the forward market exchange stuff via great conversations facilitated by great investors like Chris Fralic, who introduced us to, which was building an advertising API aggregator for advertisers (kind of one side of AppNexus maybe) before they pivoted 12390435832095842 times to whatever they are now. So our plan was forward market and that was where we would create demand for publishers (and hence revenue for ourselves), but step one was to sign up publishers because we had a value proposition for publishers pre-advertisers (this tech stack that let them create their own market), whereas we had nothing for advertisers without inventory.

Signing up publishers sucked! The problem was, in advance of the other side of the market, there was no short term revenue opportunity. They could create a market, but that market wasn’t going to move their needle this quarter or next quarter. It might nine months out, but sales had needs today. Our hypothesis that publishers would never build their own advertising sales infrastructure was correct, and they told us in customer development that it was a pain point, but it was a top ten pain point, not a top three pain point. So we signed up less publishers than I wanted, but of publishers we signed up, the actual implementation of our system was equally glacial. Our customer development had failed because we had identified an important problem, but not an urgent problem.

So I realize we have this structural customer development nightmare right as we pull together our first round of financing and I think, “This sucks!” We have two choices, double down, take on investors, drop people on site with publishers and drag publishers kicking and screaming into the market (an expense not really in the plan for the round we are raising), or take this bluebird offer to sell the company to a company that is hitting it out of the park. I sold the company! Can you blame me?

So today, I am reading about iSocket on AdExchanger and Richard Jalichandra, the new CEO, says:

As we continue to build out infrastructure here, I predict we’ll soon see rapid increases in adoption.

Today is basically five years to the date from founding Deconstruct Media. And iSocket had raised $2m before we got started (which was a pain in my ass because they raised money from what felt like everyone). (Party round as barrier to entry!)

Certainly, I felt like their adoption was glacial as we were moving, to see Richard say that made me think, “it hasn’t gotten much better”. I suspect that fundamentally while publishers are interested in selling more the way they used to be selling, that is going the way of the dodo and simply automating the previously manual way things were done isn’t going to represent the future of digital advertising.

So in one year we built the tech infrastructure, didn’t figure out how to sell that to customers, sold the infrastructure to another company, and then went and spent four years building a huge, huge business in the mobile market. Smart? Cowardly? Both? Dumb? Brave? Blonde? I guess this tells you the Hard Thing About Hard Things. (You are damn right that is an affiliate link. I am an arbitrageur and I need the money.) I don’t mean this as a criticism of iSocket, I hope they print money. If they build a big business and this industry turns out huge, then I get to tell people I was one of the first companies in this huge market. My point, maybe more than any other point, is customer development is hard. All these publishers told me how exciting this was and what a big problem it was and a bunch even signed contracts with me, but it wasn’t an urgent, burning, agonizing pain. Those are the only kind of pain worth solving as it turns out.

As it turns out, for Deconstruct Media, it made more sense to do something else than start customer development over, but sometimes you don’t have a choice. I am doing customer development again. It’s fun!

P.S. I wrote this prior to the announcement that iSocket and ShinyAds were acquired by Rubicon Project. Despite raising more than $15 million dollars, they sold for $30 million. It brings me no pleasure to realize that such an outcome was the best the market could offer after years of work.

Leave a Reply