Pleading With Your Executive Team

Problem

Not all stories have happy endings. Some of the most tragic ones that I have to tell have been great learning experiences, however. In this story, I got poached by a company that specialized in mobile app downloads.

This company used their own proprietary system that would deliver apps to the home screen of new phones acquired from carriers like Verizon. The use case: you buy your new Samsung smartphone. You fire it up for the first time. Whether you notice it or not, apps are already installed on your home screen.

The company built a demand- and supply-side marketplace to make that happen, driving apps on behalf of developers who needed users to the phones of customers who might be a good fit. The company wanted to get into programmatic ad-buying, as well, which was outside of their ecosystem at the time. This would unlock an interesting set of use cases tied to the core business -- for example, to target users who were using match 3 games on their phones and to present them with, say, an ad for an alternative match 3 game for them to download. You could also retarget users who haven’t been active recently in a mobile app, and bring them back to that app to re-engage.

When I was recruited for this, they had a version of this product infrastructure in alpha stage, very much test mode. They were able to buy traffic, but the throughput was extremely small and totaling a few dollars of revenue per day.

Actions taken

They were crawling forward little by little. I entered the scene and wore many hats along the way. Not only was I managing the product --- I was working on the company’s business development, too, seeking new mobile real-time ad exchanges where we would be able to reach more users. We wanted more access to greater volume.

I was also working closely with our Engineering team. We wanted to figure out how we could increase our ability to consume more overhead, to get past the point where we were handling only a few auctions per day. Working on data strategy and how to get smarter about our capabilities on the frontend was also in my wheelhouse. We wanted to be more deliberate in the types of customers that we were targeting with our ads, and use the data in our ecosystem to be smarter about our audiences.

This involved not only finding the right way to optimize, but also learning from the experience in order to extrapolate that insight outward. There was an opportunity to monetize that data and to sell it, but it would also be an invaluable resource for our own business, as well. We were able to improve our system continuously with the help of this insight.

Many of our clients had very low budgets. In some cases, they were looking to acquire new users for as little as a dollar. At that rate, one hundred thousand dollars would buy you one hundred thousand users, but this assumes very high levels of engagement with the ads. Imagine that For every dollar spent to acquire a user, 10 users on average clicked on the ad at an average cost of 10 cents per click, while one of those 10 actually converted.

Getting a user for one dollar is not always easy. It would only really be possible when you were working with these really big, marquee brands that the user already recognized. For the economic constraints of a $1 acquisition cost, you can generally only afford to buy “bargain-basement” ad placements, not the premium ad placements that look the best. Users need to sift through what’s in front of them in order to even see your ad in the first place, or even be on less “sexy” apps in the first place. That’s what you’re buying at that price point.

We could convert users while hitting that one-dollar target if the brand of the ad was recognizable. Most of the company’s relationships were with these smaller, less well-known brands. They were apps that the customers did not recognize right away. It was more difficult to convert them, in those cases.

Within a few months, I was able to demonstrate an incredible improvement in velocity; the dog had teeth and was able to hunt if we were using marquee brands, and he could hit or beat the $1 acquisition cost target. For our more obscure customers, however, we had concluded that we would not be able to realistically hit those numbers on a sustainable basis.

A few months into this we tried to have an intervention with our CEO. We laid the facts out before him and shared our perspective on what the data was showing -- this plan won’t work unless we have more well known brands as our advertisers. We also gave them three or four scenarios that would utilize what we had already built and enable us to do some really amazing things to pivot -- in other words, capitalize on what we’ve done to unlock a different value set, without throwing it away or starting over.

This steering committee of the chief data scientist, myself, and our business counterpart were unsuccessful in convincing our CEO of the data. Instead, he asked us to continue as we were because he thought it was the right path. There was no data to support this decision. The CEO ended up shutting the product down a few months later.

Lessons learned

• Sometimes, leaders make ego-driven decisions. Our CEO would not hear any of what we had to share; they were focused solely on their own opinion. This is a terrible way to lead.
• What is the best way to approach a leader with bad news? It can be difficult to disagree, even under less strenuous circumstances. I try to be a diplomat whenever possible. Our empirical findings were very real, which, normally, can be very convincing for some types of people.
• Some types of leaders shoot from the hip without any sort of defensible strategy. These leaders tend to chase money more than anything else. It’s a common problem in start-ups. The moves that they make never compound upon one another. Money is the only determining factor. There is no sense of prioritization.