Lesson: Product Leaders with Adam Nash

Step #5 Engagement: If you do not measure something well, it is very hard to execute against it

When I was at eBay we spent years arguing about how to measure engagement properly. The good news about eBay was it's a very easy business from a financial standpoint. It's a marketplace. We take a transaction tax, effectively. A tax on a sale. We also charged listing fees. That was kind of a mistake, but we did it anyway. We made money in a very easy way.

So it was very easy to understand how we made money from sellers. It became very hard to understand how we made money from buyers. Even though, logically, you know that if there is no buyer nothing would ever sell on eBay, there'd be no transaction. This was one of the big things that eBay figured out in the middle of the decade.

At LinkedIn we had a similar problem. Most of the LinkedIn experience has nothing to do with revenue. It has everything to do with engagement and data, so we had a lot of problems. How do you measure engagement? If you don't measure something well it's very hard to execute against it.

We spent years, like most start-ups, arguing about how to measure engagement. Every start-up that's successful, truly believes that they are a unique snowflake. Their product and service is different from every other product or service that's come before it. Therefore, they never want to use any of the metrics that any other company uses. They always want to make up their own metrics.

At LinkedIn we would argue, "Well, we don't want to use page views because, actually, business people are really busy, so if we can take a flow from five pages down to three, that's actually better for business people. So you should never prioritize page views. One by one we would argue against every engagement metric.

I'm here to tell you that when you have a lot of teams, building products and features, it's really, really useful to have standard metrics that everyone can relate to. Even better, if you use metrics that other companies use, you start to have some insight into the strengths and weaknesses of your own business.

These metrics became really popular when the Facebook platform launched. Everyone who built over F8 suddenly knew what their daily active users were and their monthly active users and knew what that ratio was, and they cared about that. I think that had data a lot of value in the start-up community, because all of a sudden people were comparing numbers with each other.

Someone would say, "Oh, we have a great number there. We have 30% daily active users over monthly active users." Of course, Facebook would come out and say, "Well, we actually have more than 50%." People were like, "Wow, I didn't even know that was possible". It made everyone raise their game in terms of what was possible.

I always like to frame things as basic user insights or business problems, so I will say that these three metrics here, are actually fundamentally business problem metrics. Using that restaurant example, "how many people visit my restaurant? versus "The total number of people who have ever visited my restaurant" is an interesting way of looking at things. It basically says, look, if this restaurant went away tomorrow, these are the people who would actually care. They'd even notice. They'd be walking by and they go like, "Wasn't there a restaurant here that we used to go to?"

Daily active users is even more dedicated. This ratio gets you like, if your restaurant closed, people would actually be upset. Those people actually were like, "Oh wow, now where do we go for lunch?" right. They've actually built something important around you.

The last one, however you measure it, whether it's page views, time on site, the metric we used at the LinkedIn platform was actually "actions". We had a dozen different actions people would do to contribute content to the system, and we looked at the number of actions per daily active user.

The great thing about these metrics is, whether you're building an iPhone app, you're implementing a search engine, you're looking at a home page for an advertising campaign, you can actually find analogs for all of these. For the first time, when we settle on these metrics, teams could talk to each other about what the strengths and weaknesses of their features. You discover that, actually, some of these numbers look fantastically good for search. The percentage of people who were daily searchers over the monthly was actually very, very good. People found LinkedIn search really, really valuable and kept coming back to do it.

iPhone app, these are very easy metrics. Total number of downloads, monthly active users. We found out that the iPhone app got much better daily use than the desktop website. It sounds obvious, but when you put things in numbers like this, it proves the point and allows you to really ask, "What are the other use cases that people want to do on a daily basis with LinkedIn? How is that different than the desktop website?"

I'm a big advocate for, as you build-out companies getting standard metrics. You can have a thousand metrics if you want, but make sure that at least a few of your top-line metrics that everyone understands are the same across the company. Better yet are things that other companies, that you can compare numbers to, actually use as well.