Numbers are the fastest way to see what is really happening with your MVP. They don’t tell you why, but they show you where to look. This guide covers the quantitative data a product founder needs, how to set it up without a data team, and how not to be fooled by it.
In one sentence
Quantitative data is what people do, counted.
It tells you what is happening, not why.
Take the 2-minute quizWhat quantitative data is
Quantitative data is anything you can count about how people use your product. In an early MVP, that usually means:
- Sign-ups: how many people ask for access or create an account.
- Completed tasks: how many do the thing the product exists for, such as creating a first project.
- Return visits: how many come back, and how soon.
- Payments: how many pay, pre-order or join a paid pilot.
What numbers are good at, and what they miss
Numbers are good at three things: showing where people drop off, checking whether a change made a difference, and comparing results against the target you set. What they can’t do is explain. If 40 people sign up and only 5 finish the first task, the numbers tell you there’s a problem. They don’t tell you whether the task is confusing, the promise was misleading or the wrong people signed up. For that you need the other half, covered in Qualitative data for your MVP.
The five numbers to start with
- Interest: sign-ups or access requests from people in your target group, divided by the visitors you sent. It shows whether your promise lands.
- Activation: the share of sign-ups who complete the key task. It shows whether people reach value, and whether the product is clear enough to get there.
- Retention: the share who come back after a few days or weeks. For most products it’s the strongest sign you solve a real problem.
- Conversion to paid: the share who pay or commit. Saying “I’d pay for this” is cheap. Paying isn’t.
- Referrals: people who bring others without being asked.
How to set up tracking in an afternoon
You don’t need a data team or a complicated tool. A basic analytics setup is enough:
- Write the question first. For example: “Do people finish the first task?” Track only what helps you answer it.
- List the steps of the key journey. Typically: visit, sign up, first task, return, pay.
- Track one event per step. Give each a clear name, such as signed_up or first_project_created.
- Test the events yourself before you launch. Broken tracking is the most common reason a test teaches you nothing.
- Look at the numbers on a fixed day each week, and write down what you see next to what you expected.
Reading a funnel
A funnel shows how many people make it from each step to the next. For example: of 100 visitors, 20 sign up, 8 finish the first task and 3 come back. Don’t look at the totals. Look for the biggest drop between two steps, because that’s where the product loses people, and where your next round of conversations should focus. (These numbers are only an illustration. What counts as good depends on your market.)
Reading small numbers
Early on you’ll have dozens of users, not thousands. That’s fine, as long as you read the numbers as direction, not proof. One person out of five is 20%, but it’s still one person. If the same problem shows up again and again, trust it. If a number moves by a few people, wait for more data before you react. I cover this in more depth in The role of data in validating your MVP.
Numbers that look good but aren’t
- Page views, downloads and followers show attention, not need.
- Total sign-ups hide whether the right people signed up and whether they did anything afterwards.
- Averages can hide a small group who love the product and a large group who leave.
When the numbers mislead
- The wrong audience. Traffic that isn’t your target group can make results look better, or worse, than they are.
- A test that is too short. A week can’t show whether people come back after a month.
- Changing things mid-test. If you change the product halfway through, you can’t tell which version produced the result.
- Missing tracking. Check that every step records correctly before you trust the funnel.
Pair numbers with conversations
The best use of numbers is to tell you who to talk to and what to ask. Spot the drop, then ask the people on either side of it what happened. Together, the two give you a decision, not just a chart.
Frequently asked questions
Do I need analytics before I launch?
Yes, at least the basic events. If you add tracking after launch, you lose the data from your first and most valuable users.
How long should I run the test?
Long enough to see people come back. For many products that means at least a few weeks. Decide the length in advance, along with your success targets.
What is a good conversion rate?
It depends on your market and product. Set your own target before launch, based on what would make the business work, and judge the result against it.
Should I run A/B tests on an MVP?
Usually not. With few users the results are too noisy to trust. Make one clear change at a time and compare before and after, then confirm what you see by talking to users.
Which events should I track?
One per step of your key journey: visit, sign-up, first task, return visit and payment. Add more only when a specific question needs them.
Building a product?
If you want a first version that gives you clear numbers, take the 2-minute quiz to get the next step that fits. Or book a call and we’ll plan what to track and what to test first.
