Tips to Decide what metrics to track
There are many things you can track every day about your business, maybe way too many things. As an eCommerce business, for example, you can look at the number of product page views generated each day, or the number “hits”, interactions with your business, but are these metrics giving you a clear picture of the health of your business?
A lot of metrics businesses track fall into the category of we can call “vanity metrics”, these are metrics are important to keep you somewhat updated, but are they telling you anything about trends in your business growth? No very much… You are looking at the trees, when you think you are seeing the forest.
There are no bad metrics, but unhelpful metric surely do exist. If your goal is to unlock opportunities for growth for your business, reporting metrics will not do it for you. They are important if you handle the day-to-day operations, but growth lies most of the time in uncharted territories.
Here are a few examples of metrics that should be used as a starting line in conducting experiments, but should always be complemented with exploratory metrics, indicators that help you discover causation between seemingly unrelated metrics.
Descriptive metrics
Theses are the metrics that tell you about the past, they can be useful for reporting, and are in fact a good place to start when it comes to troubleshooting. For example, you notice that your eCommerce website have had a surge in bounce rate, the data in itself will not help you, unless compare the bounce rate with that of another period. Maybe, you notice that sells drop off during holidays each year, so you plan accordingly, and try to understand why by conducting tests.
Vanity metrics
In analytics, focusing on sheer impressions number or even worse, the so-called “number of hit” is in most cases a useless activity, the sole exception being websites with a business model that reward them for traffic (think news websites).
In the startup world, it’s the founder who thinks his app is doing great after being featured in one of the top fill in the blank of the year, only to discover that almost no one is using your app after a few days after downloading it. Soliciting media coverage is too early if you can’t convince core users to use your app.
Correlated metrics
Let’s say we find a correlation between animal on the road during summer and people getting into road accidents. Should we jump to conclusions and get rid of these stray animals? The animals are not causing the accidents, it’s the season change.
A more reasonable measure is to add signs, in those areas where accidents are most frequent. Finding the correlation is always the first step. Next, we need to get to understand causation. Meaning, we need to find what is the most contributing factor that causes things to malfunction.
Understanding the relationship between variables using A/B testing, cohort analysis, and other similar testing techniques, is always a good idea before making conclusions.
What makes a good growth metric?
Again, when we talk about “good metrics”, we mean metrics that can helpful in identifying growth opportunities. Almost any data you gather can by definition inform you about some aspect of your business.
A helpful way to identify growth metrics, is by answering the following questions:
Is it possible to use for comparisons?
Instead of thinking about the number of engaged users, a better metric will be the growth rate of engaged users between two periods. This will allow you to 1) get a sense of how fast you are progressing, and 2) establish relationships between seemingly independent factors.
Let’s take the evolution of the launch rate of an app in relation to the number of free features on your mobile app as an example. Maybe you will notice there is a sweet spot between how much stuff you are offering without compromising your profitability.
For example, you might find that there is a sharp drop of the lunch rate of the app once the user clicks on a feature only to find out it’s paid. If the percentage of drop is very significant, what would you do in such a case?
Is the metric helping you take action TODAY?
Let’s say you are experimenting with a new feature for your app, after a few months you notice that the user are using as a workaround to do another task. Should you keep developing the same feature? Absolutely not, it’s a pretty straight forward indicator to change course, or else you are throwing product development money down the drain.
Is it simple to grasp?
Finally, the metric must be easily understood by all parties involved, especially if It’s linked to a bottleneck you’ve identified in your business model (OMMT or the one metric that matters). An eCommerce business might find that there is a sharp decrease in the conversion rate between two steps in the funnel. It’s an insight everyone can understand, which make it more likely to come up with creative experiments to improve upon it.
Using ratio metrics to track growth
Finding unexplored possibilities is not an easy process, but using metrics, especially ratios, since:
1) They are by their very nature comparative: measuring conversion rate from moth to month will allow you to detect patterns.
2) You can identify weak points: by looking at your funnel, for example, you may notice a sharp drop in conversion rate between two steps, which lead us to the third reason why ratios are great for identifying bottlenecks.
3) They help you to take action: the goal of tracking your data is to identify weak points and correct them, ratios help you with just that.