Every startup has a certain percentage of “good” users. I put good in double quotes because the definition of a good user will vary depending on the startup. You need to define what a “good” user is for your startup. Let’s look at an example…

For an early stage startup that’s got an MVP in the market, you’ll be looking at some form of engagement. In Lean Analytics we call this the Stickiness Stage. The amount of engagement depends on the product–for a consumer app you might be looking at daily or weekly engagement; for a B2B app you might be looking at weekly or monthly engagement. If your startup is highly seasonal (e.g. tax software, selling xmas ornaments), measuring engagement is more difficult, but that’s not the case for most of us.

The engagement that you care about is also up for discussion–are you looking at logins? Clicks? Page views? Other actions? Presumably you know for your startup and product, what “good use” means, and you should define that clearly. Logging in isn’t enough, especially if they just log out–so you’re looking for some level of deeper engagement, something that indicates the user is getting value from your product.

For our example, let’s go with the assumption that a “good” user should “engage” with your product at least once a week. You believe this is reasonable because of the type of product (it’s not a daily use app), the problem you’re solving, and the qualitative feedback you’ve collected from users (those that tell you they love it seem to use it once per week). You might also have enough data (a few months or more) that shows you that users who use your product once per week seem to stick around, which is a good thing. You’re not pulling “once per week” from your ass, there’s some fundamental / reasonable assumptions behind it.

You can now calculate what percentage of your users are “good”.

Ideally you have a threshold that you’re aiming for, what I would call a “line in the sand”. It’s difficult to find a benchmark, but pick something. I’m going to skip passed this a bit, but let’s say we’re aiming for 25% of our users to be “good”, or what we’ll now call “active users”. And right now we’re only at 15% active.

And btw, in this case, active users is your One Metric That Matters. It’s the single metric that everyone is focused on and working towards improving.

So what can you do to try and improve the % of active users?

Start by looking for commonalities amongst your active users.

What makes these people tick? What separates them from everyone else who isn’t using your product as frequently or who churns out quickly?

It’s in these commonalities that you may find the answer to what you can improve / change / focus on. For example, you might find that a lot of your active users share similar demographic information. That might indicate a shift in market focus is needed. Or maybe you find that a lot of your active users take a particular action (read: engage w/ your product) more times than non-active users. In this case, the particular action in question could be a leading indicator of becoming active. At minimum there might be a correlation there that you can test further.

If you do find a correlation (i.e. users that do something inside the app are more likely to stay active with the product), then you’ve found what you need to focus on for experimentation.

And the big question becomes: How can you get more users to take the action in question?

You can now come up with a bunch of experiments and:

• see if you can get more people to take the action you want; and (more importantly),
• see if that results in more active users.

If you do crack that nut you’re in a very good place. You’ll know what to focus on until you get your percentage of active users to the threshold you want. And you have a process for ongoing experimentation that keeps everyone aligned around the key problems and goals for the company.

I think a lot of startups ignore commonalities amongst their best users and as a result lose out on a lot of learning from them. You should know everything you can about your best users. Similarities amongst your best users will help you refine your target market (you can’t go after everyone!). And what makes your best users different from everyone else will help you drive more people to become good users, and help you find new users that more quickly become good.

tl;dr

Here’s a simpler description of what I just wrote:

1. Define a “good” user for your product/business (which depends on the product, business, and stage you’re at). Be aggressive (aim high!) and don’t cheat yourself. Be intellectually honest.
2. Look for commonalities amongst those “good” users (it could be anything!)
3. Figure out how to get more “good” users (could include feature experiments to encourage more/different usage of your product based on what good users do, could be a shift in market / marketing strategy, etc.)
4. Rinse and repeat.

Photo from Flickr.