Know-How Exchange : Research/Metrics

I would hope the intent is not that you would update every metric every week! That's probably not going to be very useful ... and it could even lead to some wrong conclusions.

Why not update each metric, say, quarterly ... or even annually. And report each week on the metric(s) that were updated most recently and which are expected to update next.

If you want to add even more value, you can report internals or additional analytics on weeks when there are no recent updates. Same with industry data that may impact your business.

agreeing with Michael, the first thing to figure out is exactly who is going the use the KPI to make what kind of decisions. Weekly measurement is great if the KPI is to be used every Monday morning in a meeting that will decide on direction changes based on the KPI (eg TOMA has dropped below a threshold X%, so we should adjust our marketing to increase it). If this is not the case, monthly or quarterly measurement makes much more sense because
1) customer perception KPIs typically change very slowly (unless there is a crisis)
2) they are very costly to measure right; i.e. you need a large sample representative of all potential customers to claim that an increase/decrease in TOMA is real instead of noise (measurement error)

As to how to measure, even the largest companies I work with have outsourced this to research specialists such as GfK, Kantar, Europanel, Millward Brown, Ipsos etc. In many of my published papers, I do use GfK's weekly metrics to study e.g. price and service image changes in a retailing price war (van Heerde, Gijsbrechts and Pauwels, JMR 2008), Kantar's weekly metrics on awareness, consideration and brand liking (Srinivasan, Pauwels and Vanhuele, JMR 2010) or monthly changes to the same metrics in emerging and mature countries (Pauwels, Erguncu and Yildirim, IJRM 2013). In each case, the research company has a panel of several thousands of consumers, and rotates through the panel each week so that the same respondent is only asked about the same brand every 6 months. Obviously, this is too large a task for even the largest companies to do themselves.

As to online versus offline measurement, I have started to work with the daily metrics of YouGov, who poll thousands of online respondent on brand KPIs. This is typically less expensive, but of course your sample is limited to people online. In the award-winning Marketing Science Institute Report (Pauwels and van Ewijk 2013), I show that online KPIs and offline KPIs do not substitute for each other: they contain different bits of information.

Finally, a great way to cut into the potential KPIs demanded by your company is to verify which ones lead performance (instead of change with/after performance). In my work, this has cut the KPI list from dozens, sometimes hundreds to 5 to 15 leading KPIs. See Chapter 8 in my book "It's not the Size of the Data - It's How You Use It: Smarter Marketing with Analytics and Dashboards" https://www.notsizedata.com

Hope this is helpful - don't hesitate to get in touch!

Cheers

Prof Koen Pauwels
koen.h.pauwels@gmail.com