Just like your website, it’s critical to measure the performance of your analytics dashboard.
Are people using it? Do people attain significant value from it?
This is the second post in a blog series about redesigning your analytics dashboard. Start from the beginning.
Why benchmark
So, you’ve decided that your current dashboard isn’t working anymore and you want to overhaul it. Awesome!
Before you throw the good out with the bad, first conduct a survey. Measure how many people use it, how people use it, and its impact on business decision-making.
A few months after the new dashboard launches, survey the same people again with the same questions. Then pop the champagne to celebrate your biggest gains and scrutinize where you can still do better.
Who to survey
Include everyone who currently receives the analytics dashboard, plus everyone who might be interested in receiving it and whose job is directly impacted by the website.
- Leadership
- Analytics
- Marketing
- Digital product owners
- UX + web designers
- Web writers + content publishers
- Web developers
- Business analysts
- Sales + customer service
Not everyone will be able to act on the insights in the dashboard, but everyone will benefit from a shared understanding of the site’s performance.
What to survey
Again, the purpose of the survey is to:
- Determine how many people use the dashboard
- Determine its most useful information (and the inverse: its least useful information)
- Determine what information is missing from the dashboard
- Benchmark the dashboard’s performance as a communication tool
Write out your research goals before you craft your survey, then map the survey questions back to your survey goals. If a survey question doesn’t fit your survey goals, you’re breaking one of your own rules. (Actionable data, not interesting data—right?)
Respect your co-workers’ time and keep your survey to 10 questions or less. Bonus: you’ll also get higher survey complete rates.
Ideally, responses should be anonymous. Be sure those invited to the survey know this, too.
Here are some example questions you might include in your survey:
- Do you currently receive the website dashboard? (Include a link or screenshot of it, just in case people aren’t sure to which dashboard you’re referring).
- If no, do you think the dashboard could help you do your job more effectively?
- Which aspects of the dashboard do you currently look at or use? (Suggestion: use screen captures or general language to describe the different components of your current dashboard to ensure everyone is referring to the same information.)
- What information is missing from the dashboard that could help you do your job more effectively?
- Does the current dashboard help you do your job? (Yes or no / likert scale)
- Does the current dashboard help you make more informed decisions about the website? (Yes or no / likert scale)
To prevent boredom among your survey participants, mix up the question formats:
- Yes / no
- Likert scale (e.g., on a scale of 1 to 7, how much do you agree with the following statement)
- Write-in
- Multiple choice
For multiple choice questions, include an “other” option with the ability to allow survey takers to write-in their own answers.
Before sending it out, test the survey questions with a few co-workers to ensure the questions make sense and cover the topics they care about.
Ask someone from your executive leadership team to email the survey for higher response rates.
What to do with the survey results
Once the survey results are in, identify the key findings and document your benchmark metrics.
Communicate the results of the survey to key dashboard stakeholders, plus:
- What you’re doing with the information
- Your next steps
Ideally, share the results and next steps with all survey participants to demonstrate that you are listening and acting on their feedback.
Next: Conduct stakeholder interviews
Interview each web stakeholder separately to understand how they make decisions in order to improve website performance, what information they use, and what information they wish they had.