Not everything that can be counted counts and not everything that counts can be counted.
- Albert Einstein
Adam Hodge, responsible for National Marketing and Communications at the Australian Red Cross Blood Service, had a fundamental response to my religious diatribe last month, titled Let There Be Light.
Adam deemed it “both entertaining and creative,” before he lowered the boom: "...yet to my dismay [it] offered absolutely no valuable advice or insight into solving the challenge of Web site reporting for marketing professionals."
Yes, the truth hurts.
Adam continued, "We currently provide a myriad of web effectiveness reports to our board and I am sure that 90% of them are not really that valuable. However, knowing which 10% to keep and what to discard is the hurdle I face.
“We have engaged several different external consultants to assist in this task and each and every one has a different (and seemingly logical) take on the issue,” Adam wrote. “This has resulted in what I consider to be management over-reporting. I would dearly love to read an article which compares valuable and not-so-valuable statistics...which measures are the most effective for tracking site success and which are just a waste of paper.”
First the bad news, Adam: I'm a consultant. To prove my assertion, the answer to your question is, "It depends."
The less we know, the more time we spend interpreting what meager details we can get. We started out with precious little data about what the Web server was doing. We hoped more data would be more revealing. But now, Web analytics tools and Web site review companies can crank out more reports than imaginable.
There is a wonderful story in Mark Graham Brown's book, Keeping Score: Using the Right Metrics to Drive World-Class Performance, about the new head of an overburdened financial reporting department who stopped the delivery of all reports for a week. This resulted in a very few phone calls from a small number of managers asking for a modest handful of reports.
Which reports are actually useful to your organization? It depends.
In my role as a Web metrics consultant I spend a great deal of time explaining what *can* be measured, so that the advertising and marketing department, the customer service managers, and yes, even the board can start deciding what they *should* measure. It all boils down to what they consider "success."
If you're running an ecommerce site, the goal is sales. Everything that contributes to more and better transactions should be measured. If you own an advertising-supported content site, the goal is more pages viewed by more of the sort of people your advertisers want to reach.
The Red Cross site may have many goals:
- Increase brand image
- Provide emergency information
- Recruit employees
- Solicit donations of time, talent and legal tender
The fundamental metrics that help in all of the above are the ones you already know and love:
- Number of visits
- Duration of visit
- Depth of visit
- Page views
- Registrations
- Application interaction
- Ease of navigation
- Etc.
These are the kind of reports that end up in multi-colored charts and graphs stuck to cubical walls and summarily ignored. Management over-reporting is a very common ailment.
So it's time to look at this question from the other end. How do the advertising and marketing department, the customer service managers, and yes, even the board determine whether they are successful?
Start with a specific goal, any goal, and work backward.
The manager responsible for monetary donations, for example, will set an average monthly contribution amount as the baseline. She will alter her direct mail, print and broadcast campaigns and look for an increase in the number of site visits. She will change the landing pages on the site to see if she can positively impact the number of people who click-through to the donations page. She will alter the copy on the donations page to see if she can increase the number and size of gifts.
The board is only interested in the number of visitors to your site if it results in _______. You'll have to ask them to fill in the blank. Without specific goals, there is no value in receiving the same report week in and week out. They are only interesting if there are anomalies, and exception alerts satisfy that need.
So I wander from client to client, teaching possibilities and helping identify specific goals. From there we can conclude which metrics will measure whether they are achieving those goals or not.
And Adam, as a consultant, I consider any excuse to visit Sydney a good one. Let me know if I can be of service.
