Web Metrics Demystified

kaushik.net/avinash/web-metrics-demystified/

Avinash Kaushik December 11,

Confused about Web Analytics?

How about Web Metrics?

Have you been successfully scared?

Rhetorical questions. Don't answer. :)

Let's do this one step at a time, first let's demystify web metrics. We'll do web analytics
another time.

The hardest thing you'll do in your life as a Web Gal / Guy / Marketer / Analyst / Researcher /
Jack is identify what constitutes success for you when it comes to measuring Outcomes for
your website.

Step two will be to identify the Critical Few metrics to measure those successful company,
and hopefully customer, outcomes.

So what makes a great Web Metric? And what are the factors you need to keep in mind to
ensure that your valiant efforts to measure business success in this first life will be
successful? :)

This post is inspired from my segment of the WAA's well received webinar on the newly
defined Web Metrics Standards last month.

The intelligent Ms. Angie Brown and the fabulous Mr. Jason Burby spoke about the tough
work that went into creating the new standards. I spoke about Web Metrics Demystified .

In this post you'll learn how to find diamonds in the rough, how to know that a metric you
have identified for your Management Dashboard is actually a good one, and you'll learn the
process you can, and should use, to keep your web analytics metrics relevant.

Enjoy…..

Four Attributes of Great Metrics:

Metrics are a dime a dozen. Especially on the web. There are books and blogs full of 'em.

How do you know which one is your must have darling?

1. Uncomplex

Great metrics almost always are uncomplex.

Because we did not make much headway with recommended metrics foisted on us, our
response has been to create complex metrics. Six things each with its own unique multiplier
/ variable predicting the position of the sun when visitors click on your site.

Here is the thing to think about: Decisions in companies are not made by one person. If you
want action then the democracy needs to understand performance, the democracy need to
make decisions.

Not you. Certainly not your consultant. Or best friend.

If you are the only person who understands the metric, the Key Performance Indicator, then
you have just guaranteed that your company, big or small, will not take action. Because you
know the metric, but not the business.

Don't sexify, uncomplexify. Solve for the masses making decisions. It is not as easy as you
think, try.

2. Relevant

Is the metric you have chosen relevant to your business?

Since we have so many metrics we pick our favorites and then stick with them. The problem
is that each business is unique, even businesses that seem like they might be in the same
business.

In Web Analytics: An hour A Day I use the example of Best Buy and Circuit City. You might
think that they should / would / could measure their website with similar Web Metrics.
Nothing could be farther from the truth.

The only thing they have in common is the fact that they sell large screen TV's on their
website. Everything else is different. Their business models, their priorities, how each tends
to use the web in its multi-channel portfolio.

The metrics you would use for each company to measure success would be different.

It is ok to seek inspiration from your friends and competitors. In the end truly stress test
that the metrics you identify are relevant when it comes to measuring the success
objectives that are unique to you and your website.
2/7
Remember what works for Jason might not work for Shane. And those two are close! :)

3. Timely

A few years back I was interviewing at one of the biggest companies on the web. They had
just closed their quarter, it had been tremendously profitable. I asked them what the
reasons were for that great success. The rest of this is absolutely positively 100% true.

Them: "We have just kicked off the query against our data warehouse, it typically returns the
results in three months."

Me: Stunned silence.

Granted they were a big business. But still.

I learned a very important lesson on that day. Be on time or die.

That big company's stock price is a fraction of its price at that time. While not all of it is
related to their ability to measure, you can imagine how hard it is to be successful in your
business (on the web for Pete's sake!) if it takes you three months to know what worked
three months ago.

Great metrics can be provided in a timely fashion so that your business decision makers
can….. make timely decisions.

I am not a big fan of real-time (see this post: Is Real-Time Really Relevant? ). But between real
time and three months there is a sweet spot. Find out what your sweet spot is and then
ensure your data can be collected, analyzed, and metrics provided with insights in that
sweet spot.

Even the greatest metric in the world is useless if takes nine days to get (with insights!) when
your world changes every three days (key word bids, affiliate bonuses, email promotions,
web page updates, whatever).

Be timely.

Sacrifice complexity and perfection for timeliness.

4. "Instantly Useful"

I absolutely love this one. Smooch, smooch, kiss, kiss.

I credit my early experience with ClickTracks for that love. Dr. Stephen Turner and John
Marshall had eliminated all the non value added stuff from the application so that no matter
where I went, what report I opened, it was instantly useful.

3/7
It was a combination of the fact that there were fewer metrics but also the fact that they
were presented in a way that made it easy to understand performance and get the first
blush of insights.

Instantly useful is when you understand quickly what the metric is, and you can find the first
blush of insights as soon as you look at it.

Here is a great example, the What's Changed report in ClickTracks:

Anyone can tell you what your keywords were this month, or last month. The ClickTracks
reports shows you "what you should care about", keywords that rose in their importance
this month and which ones reduced in importance.

All the complexity is "hidden", there is no crap, just stuff you should care about. In front of
you.

Now does that not simply kick butt? [Click on the image above for a higher resolution
version.]

It will take some nice analysis and time to understand all the nuances and unlock the
mysteries and deep stuff (just like say for example with your wife / girl friend, less so with
your husband / boy friend!).

But the first blush is there. As soon as you look at it.

[I think Google Analytics V2 also does this well through use of layout and color and
summaries or things like Compare to Site Average, Compare to Past etc etc. But I admit I am
greedy. Every time I look at a report, current or new, I ask for more instant usefulness! Phil
and the team humors me by making stuff even better.]

In a data democracy metrics have to meet the bar of being instantly useful. And not just
that, think of your boss, her boss, his boss. How little they know. If send them a metric and
it is not instantly useful then it will be instantly ignored.

4/7
You want instantly useful, no explanations required, because that will give you the opening
you need to show your "deep stuff", explain the nuance, highlight your analysis!

Smooch, smooch? :)

Example of a "Great Web Metric":

Let me give you a very simple example that I think will crystallize the methodology above.

I think Bounce Rate is a "great metric". Here is how it passes the required four attributes
test:

Uncomplex? Single Page View Visits. Easy to understand, explain and propagate. Enough
said. :)

Relevant? It identifies where you are wasting marketing/sales dollars and which pages stink
when it comes to delivering on the "scent". Those two things apply to most web businesses.
Bam!

Timely? Bounce rate is now standard in pretty much every web analytics tool, and available
in every report. Every day. Nice!

Instantly Useful? You can just look at it and know what needs attention, what needs to
stop. You see 25 – 30% for your site and instantly you know things are fine. You look at a
page with 50% bounce rate and you know it needs attention. You see a campaign / keyword
with 70% bounce rate and you know there is a fire.

Set aside half hour today or tomorrow or at the end of the week and apply the four
attributes test to your own important web metrics. What do you see?

Three Important Final Lessons:

Here are three lessons that are directly from the front lines (sourced from painful battle
scars!)…..

1. Perfection is…. the enemy of good enough.

Data quality on the web is not perfect, things change too fast, everyone wants a piece of
data yesterday, your competitors are strong. Don't spend time getting things perfect when it
comes to your metrics.

If you have 90% confidence in the data (how it is collected, processed, and presented) then
make a decision. Don't wait for perfection.

5/7
Too many times we spend too much time being distracted by missing tags and the hoopla
around deleted cookies and more. Follow best practices, then move on. Go for precision
and not accuracy.

As Mr. Stuart Gold says: An educated mistake is better than no action at all.

2. Critical few baby, critical few!

I owe Steve Bennett the CEO of Intuit all the credit for this important lesson.

Steve is fond of pushing everyone to identify their "critical few". Priorities. Goals. Metrics.

My interpretation of Critical Few: When the proverbial crap hits the fan what is most important .

That statement has a phenomenal clarifying power.

If your business was on the line how would you know things are going well or badly? Cutting
through all the clutter of data, what are the metrics that are your Critical Few?

Almost all of us have too many things we measure, too many things that distract us, take
away our precious time / attention.

You probably have at most three Critical Few metrics that define your existence. Do you
know what they are?

If you have 12 then you have too many.

3. Metrics life cycle process is your friend!

Metrics no matter how great have to stand the test of time. And business changes.
Repeatedly.

I recommend this simple Metrics Lifecycle Process…..

6/7
The idea is quite simple really: Use the four attributes test to identify your critical few
metrics, go measure them, then analyze the data you collect, take action. Here's the fork on
the road. If you can't action anything then perhaps it is the wrong metric for your business.
Eliminate it. If you can take action figure out how you can improve it further.

Execute the Metrics Life cycle Process in a timely manner, I recommend atleast once a
quarter.

Some metrics will stay, those are your best friends, others will outlast their value, give them
a warm hug and say bye bye.

There. Web Metrics Demystified!

Not that hard, right? Just a dash of thought, a drop of common sense and a pinch of
passion.

Ok now its your turn.

Please share your perspectives, critique, bouquets and brickbats via comments. Thank you.

[Like this post? For more posts like this please click here, if it might be of interest please check out
my book: Web Analytics: An Hour A Day.]

7/7