Author: Andrew Chen posted in Blogging on 2013-06-24 10:17:00
After launching a product, startups often ask themselves, “What next?” — and before they know it they’ve created a monster. Andrew Chen explains why you should stop overwhelming your product with shallow features and start highlighting its core value.
Editor’s Note: This guest article from Andrew Chen originally appeared on his blog as “Does Your Product Suck? Stop Adding New Features and “Zoom in” Instead on Your Product Design Approach”
Everyone’s worked on a product that’s failing despite a ton of work behind it. It’s not for lack of great ideas, or a lack of bright minds working long and hard on the product. In the startup world, often this comes because after a new product is launched, there’s a Trough of Sorrow where features are often added to try to spark traction. After a few months of this, and a few shifts in direction, it’s easy to get a Frankenstein product that tries to do too much.
At this point, adding new features won’t help — what’s broken is at the core of your product, not out on the edges. Adding more to edges won’t do anything, because most of your users aren’t even getting there.
Eric Ries has a wonderful term for what to do here, which is to consider a “zoom in pivot.” He talks about it in his book Lean Startup, as a kind of pivot you can do if your product isn’t gaining traction.
The idea of the zoom in pivot is:
A single feature in a product becomes the whole product, highlighting the value of “focus” and “minimum viable product,” delivered quickly and efficiently.
The question is, how do you pick the feature you’re going to zoom into? And how do you validate that it can work as a standalone product? And how do you execute the pivot itself and what metrics can you look at?
The actual process of picking the new product is the same as picking any new product for a startup. Ultimately, it still has to go after a huge market, it has to be differentiated against competitors, and it has to have a distribution model. You have to be passionate about it. Etc, etc. All the standard strategy issues apply, and I’ll leave this as an exercise to the reader.
In terms of tactics though, the big thing from a metrics standpoint is to try and figure out what’s actually getting enough usage to execute the “zoom in” pivot. After all, if you zoom into a smaller featureset that isn’t being used currently, that’s obviously much riskier than noticing that out of 10 features, one or two are getting all the usage, so then you dump everything else.
Based on developing a product strategy, and looking at current usage metrics, you can develop a hypothesis for what a smaller product might look like. You can also create some goals you want to hit as far as the metrics are concerned — obviously the usage of the zoomed-in feature should be much higher, but by how much? And the usage of the secondary features should become zero or minimal — are you okay with that? If so, the next step is to test it.
It should be easy to test a “zoom in” pivot — just default the navigation and the description of the product to focus on what you’re zooming into. You can even test a few ideas simultaneously if you want to.
The above suggestions focus on making the zoomed-in feature more prominent, but you can also make the other features more secondary. You can do the following:
The combination of all of the above — either by making the main feature more prominent, or the burying the secondary features — should help the goal. You can A/B test these, primarily focusing on new users, to see what the effect looks like.
From a metrics standpoint, I think as a baseline you’d want the zoomed-in feature to increase significantly in usage, and for the secondary features to go to zero or nearly so. You also want to make sure some of the aggregate stats around frequency of use, time on site, content shared, etc. to be stable depending on what you care about.
After this iteration process, picking the zoomed-in feature should be easy. You may have to go through an A/B testing process to smooth the transition from the old feature set to the minimalist one, but over some period of time you should be able to make the metrics move in the direction you want.
If it turns out the metrics are stubborn and some important metrics go down, then that’s much more problematic. It might turn out that the zoomed-in feature you picked is somehow not right enough. Or maybe the user base you’ve amassed isn’t right for the pivot. Or maybe you need to develop the feature set a bit more, in the direction you’ve pivoted, to get to the right product.
For all of these, the Plan B might be to either accept the new feature set and deal with the reduced numbers, hoping to fix them later, or alternatively, to pick a new feature set or continue iterating on the zoomed-in feature set, until it works. That’s all gray area.