Breaking Free of Abused Analytics Jargon
Analytics jargon, lingo, and terminology has been skewed over the years. I'm not able to put my finger on when this started to occur or the root cause, but terminology has drifted away from the true meaning of words and phrases to something I find very misleading and potentially dangerous to those that consume analytic services. Here are my top 5 offenders:
1. Analysis is not a report, a single data point, and/or a grouping of data tables. An analysis tells a story. Analysis should be treated like a novel and have a beginning (valuable insights), a middle (supporting findings and visualizations), and an end (actionable recommendations with implications).
All analyses should be grounded with a measurement plan that outlines the analytic approach, predicted outcomes, and the definition of success. An analysis should provide a clear path to performance improvement
2. Insights are not casual points of interests or standalone observations. Insights answer the "why?" - they must provide value to the end consumer of the analysis and be grounded by implications to the business. Insights need to lead to actionable recommendations and next steps
Not Insightful: 17% increase in Branded Paid Search spend led to a 15% increase in site traffic
Insightful: 17% increase in Branded Paid Search spend led to 10% increase in conversion rate driving total number of conversions above goal (+2,000 conversions). Reallocate spend from less effective Unbranded Paid Search to branded because branded paid search messaging A is driving maximum conversion efficiency, which is consistent language with Y creatives across X channels.
3. Goals are not designed after the fact and should not be driven solely by historical performance. Goals should be associated directly with KPIs and may be tied to frequently evaluated operational and engagement metrics. Goals should be created based on industry benchmarks, past performance, and future performance expectations.
4. Key Performance Indicators (KPIs) are not operational or engagement “front end” metrics. Operational and engagement metrics must be evaluated to understand variances in KPI performance. KPIs must ladder up to strategic business imperatives and represent indicators of success while being evaluated against goals
Not KPIs: number of emails sent, open rate, click rate (though these supporting metrics still need to be evaluated)
Potential KPIs: email conversions, email conversion rate, cost per email conversion
5. ROI cannot be evaluated while looking through a myopic channel specific lens. ROI needs to have a 360 degree view into all channels, tactics and programs. Channel specific evaluation is useful when comparing publisher 1 against publisher 2 or program A against program B but it needs to be positioned appropriately and not represented as ROI or sales contribution since it is measured in a vacuum
What can we do about it?
For Analytic Leaders: Put client needs first. It sounds obvious but often gets overlooked when analysts are heads down in their deliverables. Teach analysts the “So What, Who Cares” approach to developing their analytic story.
If content doesn't provide value to the end consumer then it doesn't belong in the analysis. Ensure analysts are using terminology accurately and help them manage expectations regarding their deliverables.
Build a structured approach that will support analyst growth while keeping their deliverables away from the analysis miscues highlighted above. Foster curiosity in analysts so their deliverables don't become stagnate.
For Consumers of Analytics: Ask matter-of-fact, analytic savvy questions to those providing analytic services. For instance: “How does this impact the business?”, “What should we do to improve performance?”, “What metrics are used for optimization?”, “What do we consider success?”.
If you find that these leading questions aren't generating appropriate answers then perhaps it's time to consider another service provider to fill your analytic needs
Mike has 11+ years of experience in analytical strategy, database marketing, and market research. His expertise is focused on market analysis and planning, customer centric marketing, segmentation, customer targeting, campaign management, research, online consumer behavior, web analytics, and ROI optimization. His analytical, managerial, and technical skills provide clients with insights into global and tactical decisions such as budgetary spend, marketing forecasts and campaign performance.
To contact Mike, email him at firstname.lastname@example.org.
About the author: