Leveraging Analytics to Achieve Product-Market Fit

This post was co-authored by .

In the competitive startup world, understanding your users can make the difference between success and failure. It allows startups to find product-market fit, grow their product, and convert users into customers.

Data builds that understanding.

That's why we created the , an initiative that gives startups access to the product intelligence and community they need to be successful. Recently we ran a Q&A session for the Amplitude community featuring , Ex-VP of Product at Acorns and current Partner at Uptech Studios, a startup studio, venture firm, and official Amplitude implementation partner. The session focused on how early-stage startups can use analytics to better understand their users.

Here, Matthew shares some key insights from the session that will teach you how to approach essential areas of startup product analytics and help you turn your users into customers. Over to Matthew.

Establishing an effective analytics plan

One of the questions we received from the community asked for advice about how to set up an analytics plan. I recommend creating a that records:

• The events and properties you're going to track
• Why you're tracking them
• Where they're going to be tracked

Your tracking plan has to work for two main roles: the developer, who has to implement the tracking, and the data consumer. You can create your tracking plan in Amplitude or an external spreadsheet. I prefer the Amplitude solution because it reflects what is live in your application.

It's not practical to track all events that happen in your app. Instead, you should track only those important for understanding user behavior or triggering marketing automation.

When defining which events to track, think about the end result. Outline what would need to happen to prove that a new feature has been successful and create a mockup chart showing what those results would look like.

For instance, almost every app needs a funnel to track registration so you can know where your users are dropping off during signup. Determine the key milestones during your product registration and include them in your tracking plan.

For most FinTech apps, registration milestones will include:

• account_created
• email_verified
• identity_verified
• bank_connected
• deposit_initiated

Approaching data ingestion

At the product market fit stage, you want to discover how customers use your app and how they're moving along different paths. We were asked how to approach data ingestion, especially since some users decline to share data with app developers.

In my opinion, you should be able to get pretty far with Amplitude without collecting any personal identifiable information (PII) from users. However, if you need to track attribution to understand where you've acquired users from, one fix is getting customers to sign up on a website before giving them access to your app.

Web signup means you can still use campaign parameters to track sources, campaigns, and keywords without prompting the user. Once you've got these parameters, you can store the properties in Amplitude. Group users into different cohorts by attribution source to identify which attract your most valuable users.

If your acquisition strategy is app-first and you don't want users to sign up on a website, you can use . The network provides accurate attribution for iOS campaigns without needing PII data and can give you a blended sense of where you're acquiring your users from.

Some users might opt into the app tracking transparency prompt and ask Apple not to track their behavior. In this case, use a service like to gather instead.

Analyzing customer journey data

Folks from a recently launched neobank wanted to know what insights they could learn from user journeys to help find product market fit. First, you need to define what it means for users to be engaged in your application. In other words: what actions show repeat usage of your product?

For a neobank, this might be tricky because some of these behaviors happen outside of that application. For this reason, spend time mapping out a complete picture of your customer's journey, including how your product interacts with everyday life outside the app.

For example, one of Uptech’s clients has a prepaid debit card product. Customers get their card, activate it in the app, and then track their transactions and balances in the app. We believed that the action that made the most sense to track as a sign of engagement was how often the user reloads their card.

However, card reloading happens outside the app at a physical location. So, we had to build a service that listens to a webhook for that event to pull that data into Amplitude.

Identifying pathways to monetization

A big challenge startups face is figuring out how to turn users into paying customers. We received a question from a team who runs a social networking app. They have over two hundred users but no paying customers, and they wanted to know how they could find the best way to monetize.

To discover which monetization options will work best for your business, set up experiments and then use Amplitude to track whether customers convert. For example, you may turn your app into a full subscription business and give every new user a 14-day trial when they sign up.

By using Amplitude, you can set up cohorts that help you capture the conversion rate of:

• How many people sign up
• How many people complete the trial
• How many people become subscribers

If you don't see the conversion that you expected, try changing a variable like price or trial length. Create a new cohort of users for people who signed up in the new experiment and track long-term retention for each cohort to see which variables are most successful.

Leveraging Amplitude for more insights

We also received a question from someone who has built a developer community where people can take classes to learn about certain topics. They're currently using Google Analytics (GA) and have found it isn't providing the data they need—they wanted to know how they could use Amplitude to gather more insights.

Back in the day, GA was great for quickly getting feedback on what was happening on your website and for marketing to track where they acquired customers. I switched to Amplitude because it gets event-driven funnel analysis right while doing everything GA does.

In the case of the developer community product, I would start by building product funnels to understand the following:

• How users land on the site
• How they find courses
• Whether they ultimately enroll

The team could add synced events for each step and then use funnel analysis to identify where people drop off. From there, they can conduct user research to investigate why at these points and inform how they should redesign these pages.

For example, one of the reasons why people might drop off could be due to them having questions about the platform. In this instance, they could try adding live chat to help coach people through the areas where they tend to drop off.

Uncovering insights with AI

During the Q&A, I was also asked to share my opinion on using AI to uncover insights. I believe AI can be a great assistant to the product manager by summarizing data and identifying anomalies they should investigate. AI can also suggest reasons why the data might have changed.

However, for now, the PM should be responsible for determining why there's been a change and reporting that back to the rest of the team.

One of the trickiest parts of analytics is getting to a point where you can definitively tell someone why a particular metric is up or down. It's not enough to define a relationship between one or two metrics; you have to fully understand everything in that particular business' ecosystem.

For Appfit, we noticed that our signups were going through the roof. We couldn't figure out what was happening: there had been no increase in ad spending that day, no new PR on the product, and referrals looked normal. It turned out that Apple had decided to feature our app without letting us know.

A similar increase happened when Appfit was featured on a morning news show without warning. In these instances, AI wouldn't be able to identify the reason for the spike in signups unless it had access to all of these other sources in real time and could understand how they related to the business.

I'm not saying that we won't get there eventually, but at the moment, I wouldn't be totally comfortable with AI telling me with confidence why my metrics have changed.

Join the startup community

If you enjoyed this discussion, you join the #startups channel in , our Amplitude Slack community. Anyone interested in accelerating your product journey can gain free premium access to Amplitude for one year by .