By Sven Kliem, Head of Pre-Sales & Customer Success, Webtrekk
Depending on your report complexity and the amount of data you are analyzing, your reports might take a while to deliver results. To avoid frustration and to make it a little easier for you to get started, we developed a little cheat sheet in the shape of this blog post. It shows you several easy tweaks to see results faster and to increase the performance of your Webtrekk Analytics calculations.
First and foremost, make sure you only include the data in your analysis that you actually want to get insights on. If your performance is still not as expected after applying all the tips below, make sure to contact the Webtrekk Support Team for further assistance.
1. Start with defining the questions you would like to get answered – and then create your reports. It will sharpen your focus and improve your report usability.
2. Build up a reporting framework for your company and share reports with all necessary users. Existing calculations for a report are delivered from the cache and hence provide faster results.
3. Are you looking for exact figures or rather for a trend? If the latter is true, you can improve analyses performance with our frontend sampling slider. Simply set it to < 100% to get quick insights. By the way: dashboards and exports are always displaying 100% raw data, regardless of the sampling slider settings.
4. Don’t set the same time frame for each analysis. We recommend you differentiate between short- and long-term analyses and set the analysis time frame accordingly. You can easily set individual analysis time frames.
a. Enter the edit mode (editing rights required)
b. Choose the report element you want to modify and click on the cogwheel
5. If you are only displaying the visualization of your analysis in the report, make sure to remove all unnecessary metrics and dimension from the data table. That way, they won’t be calculated and will not slow down the calculation process. Simply right-click on the dimension/metric that you want to remove:
6. Pivot tables can take a while to load. This is not caused by the result calculation, but rather due to the complex connections of large data sets that are inherent to a nested pivot table visualization. If you are looking to export an analysis of such type, we recommend you to change the visualization to the list view:
7. Check your account for scheduled exports that are no longer relevant. Disable those and performance will improve. This applies to scheduled dashboards and reports delivered via email as well as to raw data exports.
8. Talking about scheduled exports – be mindful of how often they need to be sent. If you have a report that only looks at annual figures, there is no need to send it daily or weekly. If the report is delivered only once a month, it will have updated numbers and be of real interest.
9. Know your account setup requirements and use the account setup that can answer your question the fastest. Webtrekk customers often have a global account as well as country or brand specific accounts. It may also make sense to have a short-term data account for a very limited amount of data sets based on the timeframe.
10. If you are looking for insights from a subset of data that is also available in a smaller account, run your analysis there instead of your global account. Using less data in this specific account will result in faster result calculation.
11. During implementation, keep the cardinality of main dimensions in mind – details for objects can always be measured through parameters and categories.
12. The number of filters and filter combinations used in an analysis increase the processing demand. Complex filter setups will naturally require a longer processing time, so make sure to only set filter required for you analytical question.