If you can bring a data driven component to some elements of what you do, you can be outrageously powerful.
Big-picture brainstorming lets startups dream up exciting products and get them off the ground, but it’s data-minded decision making that will keep them climbing upward. “If you can bring a data driven component to some elements of what you do, you can be outrageously powerful,” says Andy Johns. “As a designer who touches the product, you have tremendous influence.” That influence can be especially valuable, he says, for growth teams, the people tasked with making sure a product gets new users and hangs onto old ones. Andy always looked to the data as a product manager on user growth and engagement at Facebook, Twitter, and Quora. He shared four graphs showing how any startup can—and should—do the same.
Growth is, at its simplest, users in minus users out. To keep track of growth, companies do what’s called user accounting, tallying up how many people are starting to use the product and how many are stopping.
The details of that tally can show a data-minded team different opportunities for growth, such as:
- Increasing the rate of user acquisition, or new users signing up
- Increasing the rate of user reactivation, or previous users restarting their accounts
- Decreasing the rate of user “churn,” or current users not logging in for 30 days
Andy says one solid way to grow your product is to look for “micro-optimizations,” little fixes that, over time, give big results. “If every four weeks, you figure out how to boost your growth rate by 10%,” he says, “about 40 weeks later, you can go from 1% weekly growth to 3% weekly growth.” Making these smaller changes more often can be just as effective, and far less time consuming, than designing a new feature that ups the growth rate by 200% all at once.
A clear understanding of where your traffic and users are coming from will let you decide where to focus your growth efforts.
Let’s say 40% of your traffic comes from search engines, but you only have 0.01% conversion; that may be an area to grow.
Andy once worked on doing just that, improving a product that had a 0.2% conversion rate from search engine traffic—meaning only 2 of every 1000 visitors from a search engine joined. “We spent 14 months working on the same freaking landing page, increasing the conversion rate,” he says. But those small changes paid off: The conversion rate is now 5.5%, or 55 new users of every 1000 visitors.
Gathering the data you need means running experiments, testing out various options to determine what works best. Andy offered a few tips to make sure your experiments, and the data they generate, are sound:
- Be aware of atypical traffic drivers (like news articles mentioning your site)
- Run experiments two times, and see how they match up
- Run experiments for a full week (Mondays are different than Fridays)
- A/B testing is easier at larger companies because the data is less noisy