In my time as a product manager, I have rarely had the challenge of distinguishing “good” ideas from “bad” ones. As I get to work with consistently creative, intelligent, ambitious people, I instead find myself with the problem of too many great ideas to pursue. All the ideas seem so good — so necessary, even— and yet, there’s no way we can do them all.
What then?
Increasingly, what I try to do is to imagine what I’d show in a Superbowl Ad for my product. What magical user journey, what use cases, what features would I actually want to cover if I only had 30 seconds to show it? I often find that thinking like this leads to a surprising, but useful, sense of where we should focus our efforts.
Ranking feature requests will not tell you what your product is for
We all do it: we throw all the wonderful ideas gathered from users, colleagues, and brainstorms into a spreadsheet or backlog. Then, as clever software people, we come up with a sweet ranking function that weighs user value, complexity, revenue, and other factors. There are many, many great scoring systems, and even dedicated software, to prioritize features in this way.
There are many cases where such ranking works great. For instance, in understanding what bugs to work on, I like to rank user reports by a combination of frequency and severity.
But as a way of answering the very top level question —what is my product for? — I often find that ranking falls short.
This is because ranking is an algorithmic “black box”. It becomes hard for the PM, let alone other team members, to instinctively assess the priority of a change. Because few of us can operate a spreadsheet “in our heads,” the team’s priorities seem arbitrary. A team member with a great idea, operating from her mental model of what the product should do, might be surprised when it scores low on the PM’s 5-point scoring scale. Conversely, she may not be contributing ideas that would score very highly. It is hard, in other words, to align everyone’s “mental model” to a multidimensional scoring system.
Moreover, in a desire to get through the list of features, the tendency is to do each one “well enough” and move on. It becomes easier to do many things — but none of them brilliantly.
The SuperBowl Ad Pyramid
A few years ago at Google, I was introduced to an exercise that has since helped me come up with a top-level “litmus test” of what a product, or feature, is really for, and what’s most important to deliver.
Here’s how it works. Take all the ideas you have, and organize them into 3 categories:
The way I do this is typically:
- Start with all the ideas at the bottom (“Everything else”)
- Promote 4–6 things I would put on the “box” of my product to convince people to buy it (if it’s an app you can imagine the App Store description). These should be things that the user will find obviously useful, but interesting and distinctive features to your product.
- Promote 1 of those, in turn, to the very top of the pyramid. This is what your Superbowl Ad should be about. This thing should be epic: obviously useful, super unique, and practically magical. (If you can’t think of anything at all to put here, you actually do need more great ideas.)
For example, when we completely redesigned the Yahoo Finance app in 2013, we came up with something like this:
Our goal was to make the Yahoo Finance app the “best way to follow the companies you care about” — and in particular better than the Stocks App that ships with iOS (which, incidentally, was also powered by Yahoo data). We identified 4 key features we “put on the box”: realtime data, high quality news, great charts, and super relevant push notifications. Among these, we saw that super-relevant push notifications felt like “magic” to our testers: even Yahoos were hearing news about the company through the app’s notifications.
From pyramid to priorities: Surprising findings
Having filled in the pyramid, here’s how we got to our “minimum viable product”:
- “Superbowl Ad” thing: Invest the most in this. Make it as flawlessly executed as you can. Iterate and iterate. Don’t cut corners.
- “On the box” things: Do them very well. Make small compromises as needed, but they should still be clearly better than alternatives.
- Everything else: Default to “we’re not doing this.” Reverse that only if you really, really can be convinced they are prerequisites for understanding and regularly using the above.
What is surprising here is that a lot of features we thought were “must haves” fell to the bottom of our priorities. This included features that would have ranked highly on a feature spreadsheet because (for example) they had high usage on the existing site.
To take an example: let’s say you were designing a word processor. Your “Superbowl Ad” might be about your software’s unique ability to turn bullet points into full, grammatically complete paragraphs. So you spent a lot of time on making this perfect. In contrast, you left “Print” functionality in the Everything Else category — it wasn’t something that would really set you apart.
Most PMs would find it unthinkable to ship a word processor, even as an MVP, without the ability to print a document. But the Superbowl Ad Pyramid tells us otherwise: the core differentiators do not really depend on having it, so default to excluding it. You might find yourself in two scenarios:
- Users love the brilliant auto-composing elements from the Ad, but now they’re beating down your door to add Print. They are calling you stupid for not including it — it turns out to be essential to further adoption and use. You know you have a hit product, so you quickly add Print.
- Users don’t understand or love your core feature, and frankly aren’t asking how to print. Thankfully, you didn’t waste time building Print when the core product wasn’t resonating.
Put differently, if your product has hit features — if they would get upset to lose your product — users will give you a chance to build out those “table stakes”, threshold features. But if your product is only “table stakes”, it won’t be a hit and those won’t matter.
(This is somewhat consistent with the Kano model of product development, which tries to capture the relative benefit of working on “delighting,” “threshold” and “performance” features).
From pyramid to mental model
Finally, the pyramid provides a way to align the team on a shared “mental model” of importance.
Michael O’Leary of Ryanair often asked people coming to him with ideas, “How does this make us the low cost airline?” Sometimes the employee might say, “Oh, it won’t cost much more.” He would respond, “that’s not what I asked. It might not cost more, but it doesn’t help us become the low cost airline either. We’re not doing it”. Essentially, he was sharing a litmus test he used — and the whole company could use — to focus and evaluate priorities themselves. Low fares would no doubt be the takeaway of any Ryanair Superbowl ad.
Similarly, once the Yahoo Finance agreed we were building “the best way to follow your stocks” — and that we would do that through realtime, news, push, and charts — we could all predict what was likely to be a priority, and what we should probably not waste time discussing. Engineers came up with and executed on absolutely brilliant ideas that would probably never have come out of a prioritized list.
In short, the Superbowl Ad Pyramid has helped me identify, prioritize, and communicate the most exciting and impactful things we could be doing for a new feature or product. It is an exercise I like to do not only when first getting to MVP, but regularly for existing products – especially when it feels that user needs or the competitive market is shifting.
