Every number has a story to tell. Numbers on their own have no agenda and no bias, but, of course, the person that tells the story always does. Dumb, data-driven marketing happens when numbers have no context and people have no confidence.
A recent post over at kissmetrics outlines 4 Common Errors That Kill Data-Driven Decisions:
- Judgement, and
All excellent points, that Shayla Price, the author, provides each with solutions.
When I want to gain insight into data, I remain ever-aware of the phrase “lies, damn lies, and statistics” and that there is level of perception and subjective communication behind all communication that fuels my curiosity.
Numbers need context and transparency to interpret intent.
Too often we take numbers and data out of context and should always ask:
- Who collected the data?
- What was the sample size?
- What was the sample range (time)?
- How old is this/when was it collected?
- What was the hypothesis before collection?
- Has the hypothesis changed, to what degree and why? and
- So (now) what?
So often delivered with subjective agenda or bias. Numbers and data scare me as I need to know their role before I know the role they play.
Ask some of the above questions and level set what you are looking at as well as the way the data was prepared. Statistically, your strategy will fail, but more favorable odds are available, without going to Las Vegas.
Do not assume numbers stand equal for all. Do not let the innumerates run the asylum, you, your customers, and your business deserve better number evaluation. Spot the innumerate amongst us as those: marked by an ignorance of mathematics and the scientific approach
adjective, without a basic knowledge of mathematics and arithmetic
noun, a person lacking basic knowledge of mathematics and arithmetic
Numerical illiteracy, stands in the way as much as perception.
Show Your Work
When I deliver numbers, data, and recommendation, I try to keep in mind my 4th grade math teacher’s advice, “always show your work”. As I recently wrote, seeing not perceiving.
Show your work and save you, your team, you boss, and your company credibility when presenting or advancing an argument with spotty data.
- Baseline the number: the start number with the unit of measure and the end number with the unit of measure
- Get performance period: measure in hours, days, weeks, months, quarters
- Beware the ratio: numerator and denominator matter, rate*
- Percentage ploys: starting size, performance period comparison, compounded
- Measure by measure: measure the same units, a minute and and a day need conversion to the same measure
A quick number bias sample I provide to teams comes from Google Analytics and reveals data review, numeric interpretation, and context. You can get this from the Behavior > Site Content > All Pages report and the following columns:
- Page Views: total number of pages viewed, repeated views of a single page
- Average Time on Page: average amount of time users spent viewing a specified page or screen, or set of pages
- Bounce Rate: the percentage of single-page sessions or only sessions that start and end on that page
- Exit Rate: the percentage that were the last in the session, (number of exits)/(number of page views) for the page or set of pages. It indicates how often users exit from that page or set of pages when they view the page(s)
Individually these numbers are pointless, there is no context or comparison, put them together on sentence and they begin to provide context:
Of all our pages [Page Views], people spent an average of mm:ss [Average Time on Page] on this page and x% left this page after only viewing this single page [Bounce Rate], while y% overall this was their last page viewed [Exit Rate].
For my site, here is an example:
The Impact Analysis Template page had 183 views or 14.35% of all of my site page views, people spent an average of 1 minute and 59 seconds on this page, 89% left the site after only looking at this one page, while another 86% this was their last page.
You begin to delve in to a story to pursue with a hypothesis to test. Still missing, for better context: the time frame was this a day, a week, a month?
Risk Return Ratio
More advancement, more risk advanced in numbers. Here is a shortlist set of metrics marketers provide, track, and report:
- Subscriber: baseline number, growth per period, cumulative growth
- Lead Conversion: subscribers to marketing qualified leads (MQL), MQL to sales qualified leads (SQL), SQL to customer
- Source: – traffic to conversion
- Average Revenue per Lead/Customer
- Cost per Acquisition
- Value per Customer
- Return On Investment: (Revenue by x) – Cost divided by (Production cost of x + Distribution cost of x), cost of production could include fully loaded cost of employee time
- Try the HubSpot ROI Calculator – bookmark this free resource
- Actions: comments, views, Facebook Like, Retweet, +1
You never want your credibility questioned around executive shark tanks, a short career that makes. Dumb data decisions will accelerate this.
HubSpot’s Mike Volpe wrote The 6 Marketing Metrics & KPIs Your CEO Actually Cares About that is helpful, but keep in mind all the same advice as above for social metrics that matter to your boss.
- Customer Acquisition Cost (CAC)
- Marketing % of Customer Acquisition Cost (M%-CAC)
- Ratio of Customer Lifetime Value to CAC (LTV:CAC)
- Time to Payback CAC
- Marketing Originated Customer %
- Marketing Influenced Customer %
Dumb Data Decisions
Ask for the work to understand the collective design.
Make less dumb, data-driven decisions and make better arguments.
*postscript: A fun source to, really, learn percentages, ratios, and projections is gambling and poker, to practice and retain math concepts.
David Sklansky has a great book called Getting the Best of It on mathematics of gambling I read way back in 2002, my father said, “if they taught math this way I would have loved math and remembered all of it”.
Easy to see 7th-grader, Texas Hold’em, study groups not going over well with the Board of Education, but I believe Ms. Gilmartin could help us.