Every organization wants to set expectations for progress, especially if they revitalize their analytics ecosystem with a brand new implementation. Benchmarks are key for contextualizing data and evaluating KPI progress in reporting. However, benchmarking is only useful if the strategy is accurate. While it would be great to see a 5% increase in conversions every month, you’d better have a reason to expect that aside from wishful thinking!
Similarly to other data analysis tools, benchmarking is only useful if the strategy used to formulate them fully takes into account the complexities of underlying data. With that in mind, I’ll highlight the four most common factors that our clients often overlook when establishing a benchmarking strategy, as well as some keys to success.
1. Integrating New and Legacy Implementations
All too often, our clients – under pressure to justify their investment in a new analytics implementation – push to utilize their new data as quickly as possible in reporting. Any report worth its salt shows KPI performance over time. Therefore, in order to use existing reports, clients are forced to integrate their new data with legacy data – and use legacy benchmarks to evaluate progress. This is a classic case of comparing apples to oranges. There are bound to be so many subtle differences between the two implementations that any data analysis will be rendered meaningless, and evaluation of KPI performance is no exception.
Generally, we encourage clients to lean on their legacy implementations for at least a few months following the launch of the new implementation. This gives new initiatives time to generate enough data to stand on its own in reporting.
2. Seasonality and Other “Exogenous” Factors
Historical data is often a good starting point for determining benchmarks, but it’s also crucial to understand how that data fits into the scope of market trends, the health of the economy, and the landscape of technology. A company that sells paint, for instance, has a naturally seasonal business cycle – most people paint their houses in the summer. Leaning on the performance of competitors or industry standards is helpful in this regard, since it contextualizes your own performance according to expected market trends.
Furthermore, changes in the health of the overall economy will have an obvious effect on business objectives – don’t expect the same number of conversions you had last year if there are major organizational or industry changes.
Finally, anticipating changes in technology is important for maintaining accurate KPI expectations. The lightning-fast appearance and growth of the tablet computer is a testament to this. The ways that customers interface with your company’s digital presence is poised to change rapidly, and your benchmarks should take that into account.
3. Media Blitzes, Launches, and Other “Endogenous” Factors
The most obvious data input for benchmarks is under the hood at your own organization. The cyclical efforts of your digital marketing team will yield predictable fluctuations in many of your KPIs (If not, yikes. Contact us
). Media blitzes, promotions, and other campaigns should all be factored into your benchmarking process. This same principle applies for major updates to site content or product launches. Benchmarks should be calibrated according to the predicted impact of your organization’s internal activities.
4. Update Your Benchmarks Often!
In an ideal world, benchmarks would be configured programmatically, factoring in all of the variables discussed above, and spitting out a perfect benchmark every time. In an even more ideal world, you could measure every possible variable down to subatomic particle activity at the quantum level, perfectly predicting the future of the universe until the end of time. For now, just do your best with what you have. Leverage the relevant datasets at your disposal – along with a healthy dose of awareness and common sense – to create informed benchmarks for KPI progress. Above all else, update your benchmarks frequently.
All of the factors discussed in this post are bound to cause fluctuations in data between months, weeks, and even days. Benchmarks that are data-driven and well-calibrated will add considerable credibility and action potential to insights from data analysis.
What do you think of the business of benchmarking? What would you add to this list?