Here you shouldn’t be obsessed with the matrix

Here is a fun thing that is in product management. You get up one morning and your daily active users are less than 15 %. Your conversion rate, though? Up to 8 % customer satisfaction scores only reach a height, but your nasty rate is moving upwards. Your matrix dashboard looks like Jackson polk painting, except for paint spitmers, these are contradictory indicators that you question about everything you think you know about your product.

The thing about the matrix is ​​that they are like adolescents. Sometimes they are telling you something important. Sometimes they are just dramatic. The trick is finding out what you are before deciding that you will be sorry.

Let’s start with a basic truth: not all matrix deserves equal attention at all times. It seems clear as long as you are at a meeting where someone is getting out as the page falls for 12 seconds on time. Is that bad? May be. Or maybe you have just made your product more efficient.

According to Research From McConneys, companies that take the lead in data -powered decision -making are 23 times more likely to acquire users. Great But this is what they don’t tell you: Being a data does not mean reacting to each data point, such as it is a fire alarm.

If you were driving a car and every warning light demanded immediate action, you will never be able to find it anywhere. Some lights are more important than others. Can wait for some The same principle applies to the product matrix. Still, somehow, we have created a culture where every matriculation fluctuating crisis triggers a meeting.

Related: 5 steps to make matrix that are important to your company

Decision framework

So, when should you really panic? Here is a framework that has served me well.

First, find a combination of matriculation that tells the story. Single matrix lies. When the daily active users fall but the period of the session increases, it is not necessary to be bad. You may be pouring comfortable users while your basic audience is more busy. This can in fact be progress.

Clusters have warning symbols. In addition to growing consumer growth, the engagement of falling besides growing shading? Now you’ve got a sample of investigation. It is like medical symptoms. Headache alone does not mean anything. Headache with the sensitivity of fever and light? Time to meet a doctor.

In his early days, PayPal discovered it in a difficult way. She was obsessed with measuring consumer acquisition while losing a big picture: her fraud rate was climbing faster than the volume of their legitimate transactions. The matrix was there, but no one was looking at them.

Second, distinguish between well -known and left behind indicators. Some measurements predict the future. Others just confirm what happened already. The confusion of both is equivalent to using their rearview mirror to move forward.

Customer Support Ticket volume? This is often a well -known indicator. When it grows, something breaks. Tackle? Usually left behind. When the revenue decreases, the problem began months ago.

Netflix detected this when he saw complaints of sharing passwords before the subscriber growth. Complaints were in the coal mine. The growth stall was just an inevitable result.

Third, understand the natural fluctuations of your matriculation. Some matrix is ​​a drama queen in terms of nature. They brutally fluctuate as part of their normal behavior. Others are stable unless something is really wrong.

The e -commerce conversion rate can swing 30 % daily on traffic sources, month time and even weather samples. The same day’s dip means nothing. But if the monthly revenue of your enterprise software suddenly falls down? This is not the usual fluctuations. This is a customer jumping ship.

Spatif learned that he was initially scared of daily hearing time. It turns out, people listen to only less music on Tuesday. Once they understand the natural patterns, they can see the original irregularities.

Fourth, consider the cost of being wrong. If you neglect this metric and what happens if you are wrong? What happens if you are nervous and you are wrong?

Sometimes the cost of abuse is higher than the waiting cost. Imagine re -design the flow on your entire ship because the user’s new activation fell for a week. You spend months on the project, just to discover that the drop was seasonal.

Other times, waiting is disastrous. When security violations increase, you do not wait for the importance of statistics. You immediately. They work because the negative aspect of being wrong is true but lower than the negative aspect of slowing down.

Related: Use the measurements that are really important in your business

Metric rating

Here is my advice. Make yourself a matriculation. In the upper part, apply three to five numbers that really predict your business health. They get attention daily. Everything else? Check weekly or monthly.

More importantly, train your team to think in stories, not statistics. When someone comes to you with a matriculation panic, ask them to tell you the user’s story behind the number. What is actually happening with real people using your product?

In fact, most metric movements are noisy. The signal is rare, which is why it is so valuable when you find it. I know that the best product managers have developed a nearly intuitive feeling for which the matrix deserves attention.

Unless you prepare this artillery, remember it: Your measuring tools are, not a master. They should be aware of your decisions, do not make them your Make. Sometimes, a wise decision is to close the dashboard and talk to a real user.

Because at the end of the day, the products succeed when they solve real problems for real people. No matriculation, no matter how sophisticated, changes this basic truth.

Related: Why can backfire be focused on more focus on KPI