4 Tips for Using Product Analytics to Move Fast
Written by Misti Yang, Contributor for Lean Startup Co.

“Everyone thinks they have the right vision for their product, and that they understand their users, but the reality is they don’t.” Those might sound like harsh words, but in his role as Vice President of Product at Amplitude, Justin Bauer has witnessed enterprise companies spend countless hours and thousands of dollars to release unsuccessful products. According to Justin, larger companies often fail to create the right solutions because their seemingly bottomless resources can be counterproductive. For example, a large budget can prevent a team from making thoughtful decisions and lead to nothing more than wasted money.
So, how do you prevent your enterprise from investing in the wrong idea? To find the answer, we recently did a partner webcast with Justin, and we’ve compiled some of his best tips for using product analytics to inspire worthwhile and speedy innovation.
1. Start with the data
In Justin’s experience, without data, teams tasked with building a new product often base decisions on office politics, meaning either the most senior or highest paid person in the room will dictate the outcome. “Basically, that person kicks off this waterfall process where the company will build something that isn’t valuable to the customer because they didn’t have the right understanding of what the customer wants, and it wastes months, if not years, to get there,” he says.
The first step to ensure that corporate culture does not undermine the process is to gather data. Justin recommends first exploring the numbers to understand the problem you’re thinking about solving and then releasing something quickly to measure the impact. A small success can be the beginning of a sustained commitment to data-driven learning.
Justin advises, “Find that win that then you can use to convince the rest of the organization, ‘This should be something we should invest in.’ From there, I think you can build out a potential business case as to why you need something like a product analytic solution such that the whole company could do [the same thing].”
2. Find your magic number
Looking at any and every piece of data will not prove helpful, however. “In the end, what I recommend is one [number]. Figure out what metric correlates with your [desired] business outcome,” Justin suggests. “And the reality is, I can’t tell you what that metric will be for your company. It actually really depends on your customers.”
Here are a few steps Justin recommends for finding your most important metric:
- Determine your critical events. According to Justin, those are activities that deliver value to your customers. “Examples of critical events for our customers might be things like completing a purchase, adding us onto a list, or running a query,” he says.
- Understand how critical events influence business outcomes, such as revenue or engagement.
- Track the critical event that most positively influences desired business outcomes.
For example, Justin shares that at Amplitude they have found that the number of weekly active users correlates highly with upsells. “It makes it really easy to then justify [looking] at our weekly active users because we see if we have more of them within an account, the likelihood of that account upselling is higher,” he says. If you are a business-to-business company, Justin recommends tracking your Net Promoter Score (NPS), but every company should track critical events.
3. Become obsessed with your customers (in a healthy way)
Once you can measure your business against critical events, you can (and should) become customer obsessed, Justin says. Do everything you can to understand how new products and features influence customers, with the goal of only building things that increase desired outcomes. For example, if you release a new product feature, ask: What percentage of your users are actually using the new feature, and how did that number influence your ultimate business outcome?
Another key step is decentralizing decision-making and zeroing in on customer problems, not requests. Justin says having the people who are close to the customer make decisions is “really, really critical.” This also means ensuring that the product team is connected to your customers. “[The product team] can’t be outsourcing that to the analytics department because then [product] cycles are going to take weeks before they get to the next level insight and that just continues to slow them down,” he notes.
The product team should also ask customers about their problems — not for a list of feature requests, because customers may not know the best fix for their pain points. As Justin explains: “Talk to the customer to understand their problem. That’s the first thing — not what they ask, but their problem. Then, build something that addresses that problem, and then see if customers actually use it.”
4. Develop 10X vision. Deliver 10% iteration
Inevitably, as the data pours in and your customers reveal a laundry list of troubles, your company will dream up big solutions. This is what Justin calls 10X vision. “You need to understand what are the core problems that you want to solve for, but then a lot of times people will try to build this massive product to hit that vision. And it takes a year for it to get out. We say, ‘No, think about the 10 percent thing that you could build to validate it if that is the right vision,’” Justin advises.
You release the 10 percent solution to see if your customers use and like it. If not, you build something new, measure, and learn again. Keep pursuing your 10X vision with 10 percent iterations.
Even with hefty budgets, deep talent pools, and considerable brand recognition, enterprise companies often find it hard to compete with startups when it comes to developing the next big idea. But Justin’s experience has demonstrated that corporations can excel when they combine a lofty vision with laser focus. If you want to build better products, start by getting closer to the data.
To learn more about implementing Lean Startup methodologies in enterprise companies, join us at Lean Startup Week October 30th — November 5th in San Francisco. You can watch the full webcast of the above conversation here.
If you enjoyed reading this, click on the clap button so others can find this post.

