Data lies at the heart of product management and is critical to every decision you make.
Whether you're an executive looking to bring "data-driven" into your company's mission or a data analyst who wants to move up the ranks, becoming a data-driven product manager is something that all of us should aim for.
Being data-driven applies whether you are working in B2B or B2C, for startups or large companies, or as a product manager on the first release.
Here are five ways to become data-driven for product success.
1) Formulate your hypothesis with "data" in mind.
If you're running multiple experiments at once, but you haven't taken the time to think about what you're trying to learn, then you're not in a position to be data-driven in your product decisions.
You'll likely end up using the same metrics for all of them, which will make it much more challenging to recognize where each experiment is succeeding or failing, and therefore will make it tough to prioritize your work appropriately.
Instead, start by formally stating the problem, and make sure that you've designed your hypothesis to answer a business question or test a specific assumption:
e.g., "Will users who see product X report more positive sentiment about it than those seeing product Y?".
2) Prioritize metrics based on impact and availability.
If you're already using data to identify your top-performing experiments, then what's next?
Metrics.
Unfortunately, there are a lot of metrics we can collect for our products, and many of them will be necessary — that is why you should always prioritize the ones that matter most.
Start by identifying how these metrics will affect business outcomes like revenue or growth.
This should be your starting point as you design the experiments themselves: What data can you collect that will let you identify if they've been successful?
If the answer isn't immediately obvious, then chances are good there's a more impactful or readily available experiment to run first!
An example of metrics is:
# of sessions and user engagement and retention.
These metrics will help you determine if users are sticking around, engaging with your product, and benefiting from it.
Another example is if they are using your product or not.
If not, then try to figure out why.
Product companies have a well-defined strategy for metrics in that they choose metrics that show how many users use the product and if they are using it for a longer time.
For example, if your product is a video streaming website, then user satisfaction and retention metrics will help you use your product more often.
3) Optimize for speed and accuracy.
As a product manager, you can't be expected to collect and analyze all the data yourself.
If your team doesn't have enough people on board for this, look into an analytics tool that's right for your business.
You'll get automatic reports and insights based on your goals and KPIs (which saves time), and the software will also handle the data collection for you (which saves money).
For example, say you've decided to measure engagement by how many users fill out a survey after using the core functionality.
You'll set up the analytics with this goal in mind; then, you won't need anyone to go around and manually ask people if they were satisfied - they will do it themselves!
Other than time and money, there's also information accuracy.
Analytics software like Google Analytics and Kissmetrics will be on the lookout for and report errors, and then you can ask your team to check those out - especially if they seem to happen more often.
4) Gain context to make decisions.
If you do not see any trends across your data, it's hard to draw meaningful conclusions from individual metrics or experiments.
To improve this situation, configure your analytics software to automatically show off all relevant charts and graphs.
You'll start to identify trends and look for patterns that reveal the bigger picture behind your data.
That way, you can make smarter decisions about prioritizing your work and communicating with others on the team.
An example of context is how a product performs in different countries and continents, determined by geographical data.
Another example of context is how users are using your product. For example, you may find out that most people purchase mobile instead of desktop or laptop computers.
You can then look at how those numbers change over time, perhaps correlating them to new features or other updates you make.
5) Share, discuss, and deliberate.
In a transparent product culture, it's essential to share the metrics and interpret what they mean, their relationship to other important metrics, and any action items they inspire.
This lets everyone on the team understand exactly what you're prioritizing (or de-prioritizing) while also increasing morale by encouraging an open exchange of ideas.
It also helps spread the knowledge around the company about which metrics matter most, so people can make the most of their time.
For Example:
You can share the number of sessions on your website and what they mean.
Analyze that data with your team to see a correlation between sessions and user retention or engagement.
If so, then you'll want to focus on increasing the session count while also focusing on a better user experience for users!
Conclusion:
No matter what size your organization is or how mature you are as a company, using data to make decisions about your product will always be necessary. It will help you decide on what features to build, where to spend marketing money, and how to grow revenue.
With these tips in mind, you can keep improving your approach and stay ahead of the competition!
