Alright. We’ve done the broad-stroke overview of performance optimization. We’ve established some objectives. We took a side-road to learn a bit about picking something to measure on our website. Now it’s time to benchmark.
The goal of a benchmark study is to figure out how your site is performing before you do anything at all. If done well, it should give you a good base from which to try and test different site changes (optimizations, hopefully). So you’re going to get two things from your benchmark study:
- A ruler by which to measure performance of your website (so you know if that big increase really is great or if it’s pretty typical)
- Some general insight into the patterns of use on your site (what pages are people going to often? what search terms are they using to find your site? Are they converting from search terms? How many visits to your site is typical? Whatever KPIs you are using).
Doing your Benchmark
This part can be a bit dreary but hopefully it will generate some good insights. You can, of course, hire out your benchmark study (nudge nudge).
Determine how often you will be making improvements to your site, let this be the period (length of time between measurements) you measure. If you aren’t sure, go with a week or two weeks. You will be setting aside time each period to run these data-gathering exercises, so resist the temptation to do it every day.
- Locate the reports in your analytics package that relate to the KPIs you determined back in Part 2.
- Open a spreadsheet program and make a series of columns, one for each KPI.
- Then make rows be periods (for example: week ending, month ending).
- Start at the oldest period for which your analytics package has data and start filling in the rows and columns. Don’t worry if you just installed your analytics and don’t have any historical data, just start now and go for a month or so before making any changes to the site.
- If you want to get fancy, you can run averages or medians on the data so you can figure out what is “typical,” what is “average,” and more importantly what is outside of the usual range. Anything you can do that provides context for all of your data gathering efforts is worthwhile.
To help make this a little less painful I made a very rudimentary benchmark spreadsheet over in Google Docs. Since every business has different goals and objectives (different KPIs) and different amounts of resources (length of period) it would be hard to do much more than a super-generic one that sends you down the right path. But the spreadsheet should get you started.
One thing I encourage is that while you’re frantically trying to do all that data gathering keep an eye out for anything that looks informative. Notice which pages are getting lots of visits. Which pages have low or high bounce rates. Even if these aren’t part of your KPI set, doing the benchmark might help you refine your choices for KPIs. Listen to your intuition in these instances, then track it. Stick to your KPIs but make a note of which other metrics you may want to track in the future.
I entered all the benchmark data, now what?
Well you’ve survived one of the largest barriers to getting into a process of improving your website, so you should feel pretty good about that. Once you’re done feeling good about that, notice two distinct benefits of all that arduous labor:
- When one of your KPIs is reported, you have a way of knowing if the result is out of the ordinary or not (and maybe even how much out of the ordinary). You have context for your data based on past-performance.
- You have a clearer understanding of how visitors use your site, maybe this is leading you to add some metrics to your KPI list.
With these two pieces of information, your future analysis will be more informed and your insights will be more valuable.
Quick tips to help you survive your benchmark study
- Identify the reports that matter to you (your KPI reports) and add them to the Dashboard. This will save lots of hunting.
- Save your benchmark spreadsheet somewhere that you don’t lose it, this is valuable and potentially sensitive information.
- Go period-by-period (rows), filling in each KPI for the time period. Do not go KPI by KPI (columns). You want to become aware of the relationships of your KPIs to one another and this will happen by entering them in relation to their time period. Also, switching the date range will take up a lot more time if you go by columns.
Next time in the ongoing series of real estate site optimization:
Observe and Orient: Preparing for optimization