Lesson: Product Leaders with Adam Nash

Step #4 Prioritization: With agile processes, great product releases intersperse the three priorities

Prioritization. Having been in engineering and product now for almost 20 years I can tell you this fight is never fun. I wish I could tell you there's some path that you will enjoy prioritization. Prioritization meetings, debates, arguments, they're just soul sapping, it is just so hard. You get a lot of smart people, you all agree you're doing a company together, you're going to change the world. The minute you have to agree on the fact that we can only do three things, it all breaks loose. You don't trust anyone anymore because, "How could they not agree with this, that and whatever?"

Getting a good framework around this so that smart people can thoughtfully disagree and then come to a solution and execute, is incredibly important. This is the one that I came up with and it was a product of my experience at eBay. eBay was a very metrics oriented company. It may still be. I just gave a talk there a few months ago. A lot of this is actually a reaction to what I felt eBay got wrong. Which is that eBay was incredibly metrics oriented. They got this first bucket correct.

That's a lot of the reason why eBay is immensely successful. It turns out that every business has numbers that pay the bills. It may not actually revenue numbers, they could be engagement numbers. If you're a start-up, there's a certain number that you have to hit to get that next round of funding. If you're a profitable company there's a number you have to hit to beat your competitors, or to make your business model work or to be competitive in some way.

I guarantee you there's no business out there where there aren't key numbers to track, that pay the bills. Organizations that don't learn how to prioritize metrics movers, they fail, eventually. It turns out that stuff doesn't happen by accident. The market is very competitive, there's a lot of people going after the same customers, and even if you were brilliant and came up with a unique idea. There's nothing in software that's truly protectable. The system changes too fast. Organizations that win over time get really good at prioritizing metrics movers, and the ones that don't, eventually fail.

The problem with metrics movers is it turns out they are completely and utterly uninspiring. Outside the inside baseball maybe here at Stanford and Silicon Valley, B2C circles, none of your customers actually cares how you move that number. That is a "Fortune" article that no one wants to read. Things that move numbers are things like optimizing your on-boarding flow, or improving your gmail import. You A/B tested 20 different emails and got one that converts better. No one wants to read that story, no one cares.

The problem that happens is, if you are a big growing company and you're moving the numbers, customer requests start to come in. It turns out that there is this ugly fact about customer requests. Almost none of them move the numbers. A lot of the things that your best customers want will not move the numbers. I know you are saying, "No, no, no Adam, actually there's a lot of insight from what the customers request, and actually if you listen to them carefully, things to move the needle will be there. You can pick the customer requests that move the numbers."

Every engineer wants a single heuristic. Some way of collapsing everything into a single linear prioritized list. They want a list of all the ideas, 1 through 100, we line them up, we draw the line of what we can resource, and that's it. We do everything above the line and nothing below the line. Everyone's dream.

That's how eBay ran. I don't know if they still run, but that's how they ran. I'm telling you it doesn't work. The reason is it turns out that a lot of the customer requests don't move the numbers, but if you don't implement customer requests on a regular basis, your customers will start to hate you.

It's actually very simple to understand why. It's because customers are human beings, at least, still today. We haven't made businesses for robots – maybe soon. Customers are human beings and it turns out human beings like to give their opinion to other human beings, including companies. If someone doesn't listen to you, you start to not like them. It's just a fact. We are all vain. We think our opinions really matter. The smarter you are the more you think your opinion matters.

If you tell someone your opinion and they don't listen to you, you start to dislike them. If you tell someone your opinion and they don't listen to you, and then they are super successful, you hate them. You gave advice to a friend, they didn't listen to you and then they went and made a bazillion dollars, you really don't like that person so much. We don't like it when people don't listen to us and we get even more upset when they're successful.

eBay in the middle of last decade fell a bit into this trap. While they were so good at metrics movers they weren't listening to all the customers requests and a segment of their customers felt like they weren't being listened to. Every time eBay beat their numbers for a quarter, that group got angrier and angrier because eBay wasn't listening to them. They must be cooking the books. They must be doing something, there must be some ulterior motive. There's something going on because they can't hold in their head maybe their suggestions wouldn't actually move the numbers the right way.

After that experience I became a big believer in the fact that great businesses in the consumer world that are customer centric. You always budget some time to implement customer requests. You can still prioritize them, based on frequency or impact, but you have to make it part of your process to do that, if you want your customers to be happy.

There is this instinct of, "Can't you have all three?" There must be some set of features that will move the numbers, that customers want and will delight them. I'm not saying it's impossible. Lightning does strike, you can find examples of this. It's very, very rare in my experience. Very often the delight features don't come through customer request channels and don't move the numbers.

Very often customer requests won't move the numbers and won't delight anyone, except the one person who asked for it. Maybe the one out of a hundred. A lot of the customer requests are actually complaints, so they just don't trigger delight when you fix them. It's like fixing a problem, instead of inspiring them with something new and impossible.

Basically, the reason I recommend that folks think in these different buckets is, it keeps you honest. It doesn't let you lie to yourself. A lot of smart people will rationalize that by moving the numbers, they must be delighting their customers. "They must love us. Look at the numbers, up and to the right." It's not true.

There's a kernel of truth to it, that more customers you touch, the more opportunity there is to add value. People do respond to value, but emotion is so much more than value. So I recommend that folks ask hard questions, about whether they're responding to customer requests and delight features, on an ongoing basis.