Rubric Scale for Web Analytics Analysis
G Scott Stukey
Analyst | Measurement & Analytics : New York
topics: Strategy, Search, Barbarian, Measurement and Analytics, and Social Media
filed under: Measurement and Analytics and Strategy
http://tbg.cm/p5XA8M
After several years in the web analytics industry, it pains me to see the quality of some people’s “analysis” (if you can call it that). Informing your client/boss/stakeholder that your brand has 11,000 fans is NOT ANALYSIS. Hell, even telling them that is 10% more than last month is not really analysis. The key to true analysis is to add Context, Insight, and Action Items (& the repercussions of not taking said action).
I don’t care if the report goes to the President of the US or
Governor Johnson’s Neighbor’s Dog – make sure that you go that extra step and keep shoveling away at that data until you get the statement to where you want it to be.
Governor Johnson’s Neighbor’s Dog – make sure that you go that extra step and keep shoveling away at that data until you get the statement to where you want it to be.
I’ve attempted to come up with a Rubric scale for comments & analysis written in reports. It’s a scoresheet of sorts, used to measure the usefulness of any type of analysis one does. While this is by no means perfect, or as straight forward as it appears below, I wanted to have this documented to remind myself to keep pushing.
- fail (0): A single snapshot of a metric – “our brand has 11,000 fans“
- poor (1): A change over time – “our brand has 11,000 fans, up 10% from 10,000 fans last month.“
- average (2): Including competetive context – “our brand has 11,000 fans, up from 10,000 fans last month. In the same time frame, our competitor went up only 5% from 5,000 to 5,250.“
- good (3): Include the ‘why’ – “our brand has 11,000 fans, up from 10,000 fans last month. In the same time frame, our competitor went up only 5% from 5,000 to 5,250. This was due to the paid campaign traffic driver we were running and free coupon promotional giveaway“
- better (4): Adding the ‘what to do next’ – “our brand has 11,000 fans, up from 10,000 fans last month. In the same time frame, our competitor went up only 5% from 5,000 to 5,250. This growth was due to the paid campaign traffic driver and free coupon. Due to the value-per-dollar, we recommend paid ads for future audience growth as it was more effective than the free coupon (have data backup for why we recommend paid ads over coupon; etc)“
- best (5): Informing “what will happen if we do what’s next” – “our brand has 11,000 fans, up from 10,000 fans last month. In the same time frame, our competitor went up only 5% from 5,000 to 5,250. This growth was due to the paid campaign traffic driver and free coupon. Due to the value-per-dollar, we recommend paid ads for future audience growth as it was more effective than the free coupon (have data backup for why we recommend paid ads over coupon; etc). By investing the revenue we lose from the coupon into our paid ad camapign, we can expect an increase of $1,000,000 to our bottom line.“
Now, is it always possible to have multiple comments rated “best” for every weekly report that will change the site from complete crap to 24k gold? No! Or at least I don’t care to spend 24/7 working. But you’ll always be able to find some aspect of you site/campaign/project that you can improve!
Remember to always be thinking of “If I were the person receiving this report, what does this data tell me to do?”. That line of thinking will turn the comments from “observations” to “actions”!
Shoot your comments to @GScottTweets
#FoodForThought.
