Visualize your product analytics metrics

1. Select the right output format for each of your metrics.

You can (and should) have different formats for different metrics, as the actions to be taken will be different. Some examples of output formats: A daily email with 2-3 metrics compared to the previous day, week and month. A dashboard in a BI tool (Tableau, PowerBI) with multiple metrics in different charts. A Google Sheet with several metrics and some charts. A CSV export that is available on an FTP server every 3 hours for people to download. A custom website with interactive charts and advanced analytics solutions like algorithmic detection of anomalies in each metric.

2. Select the colors you will use according to the best fit for the dimensions used in your chart.

The more colors you use (especially for decoration), the more confusing it will be for consumers. Stick to using colors for categorization and information, not for making the viz look pretty. Good color choices highlight the dimensions or groupings used in the chart; for example using red and blue to differentiate 2020 and 2021 or using 5-6 different colors to show different categories of products.

3. Use length over areas/volumes to make comparing different amounts easier.

The human eye has a lot of trouble comparing areas/volumes over comparing lines or bars/columns. This is why you should favor using length for showing different amounts. For example, instead of just plotting two series of bars closely together, maximize the y-axis range a little and change to data points or lines.

4. Align whole numbers in any tables flush right to make it easier to read.

The eye is trained to read numbers from left to right and if numbers are centered or left-aligned in a column, it’s hard to distinguish larger from smaller ones.

5. Choose the right chart type for your metrics by asking yourself what you want to emphasize.

Choosing the right chart type is crucial for conveying the right information and allowing consumers to easily understand what they are seeing. There are many guides, but this one shows a nice flow of choices that you can apply to almost any data: Source:

6. Apply the y-axis values in a natural way by not manually distorting the minimum and maximum values unnecessarily

You should never deliberately mislead someone looking at your visualization into thinking something is better/worse than it is factually. The values here are identical, but the left chart has been bounded so that the lines appear more erratic.