Metric-ocracy: Less Data, More Insight

The start of a new year is a popular time for resolutions. Many people resolve to eat better (read: ‘eat less’). Or exercise more. You know the drill.

And right about now, three weeks into a new year, is when the vast majority of those resolutions start to break down.

Of course, even among marketing cognoscenti new year’s resolutions rarely involve analytics and key performance indicators (KPIs) — so I’d like to take this opportunity to outline a plan you can follow to slim down and shape up in 2012, metrics-wise.

What Are Metrics For?

Now, in many ways there’s a direct correlation between the everyman’s resolutions (and  subsequent frequent failures) and those you can practice for marketing metrics. When we talk about “eating less” or “exercising more” what we really mean is practicing those activities in the overall context of “being more fit”. Let’s face it, isn’t that the real goal? If you actually end the year healthier than you started, would it matter nearly so much if you’d also eaten more? The purpose of a metric is to support incremental improvement toward a goal.

When companies want to get more out of their numbers they often suffer from what might best be described as “number glut”. They figure if they measure more, they will have more data (ok, true enough) and that more data means better insight (dubious, at best). In fact, I’m going to assert that the more metrics you add to your arsenal of KPIs, the more likely you are to be confused as to your true current state of business health, possibly even unconsciously. In such a case, you don’t have metrics, you have a fetish.

Let me suggest a plan for improvement that can be part of your professional new year’s resolutions.

Step 1: Identify The Real Goal

Back away from any metrics you currently look at and instead ask yourself what you want the end state to be. Make it concrete. Some common retail or lead gen goals might be: “This time next year I want to lift conversions on my site by 25%” or “I want us to double sales in 2012” or “I want to halve the number of unsubscribes from our email list”.  For simplicity, I’ll chose  one — “increase conversions by 25%” — for the remainder of this column.

While macro goals are different for each business, there’s nevertheless a rather finite number of them. Therefore, explicitly stating a primary goal might feel obvious — until you find yourself listing two, three or even more such goals, some of which might be mutually exclusive. Simple and “obvious” exercises like this help define which goals are more important than others.

Step 2: Lay It All Out

Look at all the metrics, KPIs, and other data streams bearing down on you every day, week, month.  For each one, describe how it should change if it’s moving you toward your goal. You might even choose to add some conditions just to keep yourself honest. Here are some examples:

• “Gross revenue” — Revenue can be a great measure of conversion improvement, and certainly both should rise at the same time. Just be sure to question yourself as to your goal: was it really “increase conversion” or was it “increase revenues”? If the latter go back to step 1.
• “Conversion rate” — This one is almost a no-brainer. If conversions are going up, then the conversion rate will go up, too. But the opposite isn’t necessarily true — you can improve conversion rate without necessarily improving total conversions (see the following)
• “Visitors per weeks” — Normally, you want visitors rising to get more conversions, but the caveat is that you have to be driving more of the right sort of traffic for this also to be reflected in increased conversions. If you sell red sweaters, and you drive more visitors from south Florida to your site, we can agree we’re unlikely to see a huge spike in conversions. Or you may drive too specific a niche market to the site — those red sweaters might be mighty popular to folks in Nome, Alaska, but are there enough such customers to keep your business alive?

When I said “lay it all out”, I mean it. I use 3×5 cards for this and lay it out on a table. Old school, yes, but fast and effective.

Step 3: Pick Your Favorite Date For Saturday Night

Now pick out the metrics you see as most important. I’ll let you decide what’s constitutes “important”, but personally, I ask myself “if I could have one (and only one) of these metrics, which one would I pick? What can’t I live without?” Then, “if I could only have two, which two would I pick?” (doesn’t have to include the single metric answer). This is when you’ll be glad you used the 3×5 cards. You can include as many of these as you’d like to list out, but I highly recommend you keep this to a manageable list, say 10-12 (max) of your most important metrics. If you only come up with 4, then so be it.

Step 4: Insist On Delivery

Once you’ve got your prioritized metrics listed — and remember you’ve already pre-described how each of these feeds into the goal from Step 1 — you now want to hold them to account for producing results. For the next 90 days, measure how each of these KPIs changed with respect to your goal. And make changes to your online efforts with techniques designed to move these metrics in the direction you’ve already outlined. In our example, driving better qualified traffic to the site is expected to lead to an improved conversion rate, so ask how you can drive better qualified traffic — more compelling and relevant copy to improve organic search? Goal-focused paid search ads? Paying more for specific search terms? Or, perhaps better use of negative match terms to weed out unwanted traffic? You already know a boatload of tactics for driving traffic, but which specific tactics can drive strategic improvement of each metric?

Step 5: Vote Someone Off The Island

At the end of the 90 days, list out the same metrics from Step 2, but this time prioritize them by the relative “bang for the buck” you got from each one.

Some of your KPIs will have done quite well, yielding you the most insight toward improvement of the goal. Keep those little nuggets.

But some of your KPIs are likely not to have performed as expected. If so, it’s time for someone to leave the island. Whoever is last on the performance list, well, it’s best that you part company and make room for someone new. Remove it from the list.

It may well be that a particular metric in such a sad state is one which you’ve always thought about as “important” or is perhaps considered sacrosanct in your industry — yet if this metric performs poorly in moving you toward your goal, then it is as dangerous to your new year’s resolution as a 24-hour Krispy Kreme donut shop. Drive on by!

Keep the other metrics; and, to replace the metric voted off, plug in one of your other metrics that didn’t quite make it onto your list in step 3. What we want to do is create a meritocracy — or better yet, a “metric-ocracy” — where only those KPIs who actually prove themselves effective at moving you towards your goal stay in the mix.

Now that you’ve gotten through the first 90 days, the exercise will become progressively easier. Repeat it again. And again. You’ll be amazed where you get a year from now when you’ve finished your fourth iteration.

Good luck!

P.S. Don’t combine Step 2 and Step 3 at the same time, because you think you’ll save time. I guarantee you’re doing yourself a disservice. Step 2 is about left brain specifics, whereas Step 3 is about your right-brain instincts. Good business people use both sides of their brain, but get the best outcomes when done separately.

P.P.S. Let me pre-answer the obvious question: What to do with a metric that has been voted off? Kill it forever?  I’d answer “absolutely not.”  I’d keep it available and every once in a while mix it back in to see if a second chance is merited.