Occam's Razor by Avinash Kaushik

Most web analytics practitioners define Conversion Rate as the percent of site visitors who do something the company who owns the website wants them to do. So submit a order, sign up for a email, send a lead etc.

Measuring Conversion Rate is usually the cornerstone, if not the king queen and jester of the court, of any web analytics program. It is perhaps the very first “KPI” that is measured by any good analyst and we can’t seem to get enough of it.  We report it up and down the chain of command and it occupies a place of prime importance when we present to senior management.

After all why would you not, the logic is: your conversion rate is 2.2% (shop.org study Q2 2006 for retail sites) and if you improve it by 0.01% you will make a million bucks (replace that with a really high number for your company).

Yet there is perhaps there is no other single metric that is abused as much as conversion rate, none that is perhaps more detrimental to solving for a holistic customer experience on the website because of the company behavior it drives.

The behavior most observed is that people obsess about conversion rate, it is every where, massive amounts of energies are spent on getting it up a smidgen, entire site strategies exist simply to move the number up at the cost of caring about anything else.

My recommendation is that you should stop obsessing about conversion rate, put Overall Site Conversion Rate in some appendix of your weekly / monthly presentations but that is about it.

The core reasons for this recommendation are:

• Minor Reason: Overall site conversion rate (non-segmented) for your site is a nice to know metric but it is quite meaningless in terms of its ability to truly communicate actionable insights. You got 100 visitors, 2.2 of those converted, where do you start to look for what the heck happened?
• Major Reason: A minority of visitors that come to any website come to buy (minority defined as usually less than 50% of site traffic). Consider amazon.com or ebay.com or intuit.com or store.apple.com, all major “ecommerce sites”. So if a minority of people come to your website to buy why should we obsess about Conversion Rate? Are we not guilty of letting this relentless focus on conversion rate result in criminal negligence towards those other website customer segments?

   
  The question is: If you solve for conversion rate are you solving for all your traffic and are you improving the website experience for all your customers?

My answer to that last question is a big whopping thumper of a No. We are guilty of causing sub-optimal customer focus by using conversion rate and web analytics, and I recommend that we stop that.

So what might you be missing if you only focus on conversion rate?

• Customers who will come to your website to “research”, product specific. They will never buy from your website, maybe if you do a awesomely kick butt job on the website maybe you will convert them but it is highly unlikely. An example of this is you/me using amazon to primarily read customer reviews or watch the new Bill Maher show, or wanting to read specs or print web pages with product information.
• Customers who will come to your website to “learn” about you, the company. They are looking for jobs, they are looking for press releases, they are looking for your company founders bio, they are looking for why you exist, they want your blog, they want to unsubscribe from your emails etc.
• Customers who come to our websites for “help”. This is people looking for support or looking for driving directions to your office or they want to send you a nasty email or register their product etc.
• Customers who come for reasons that we don’t know simply because we simply never bothered to ask (this is huge by the way).

All these segments, and there may be different ones for your website, will never buy from you. All these segments are not static, they change based on market conditions, based on actions your company takes (say campaigns or branding etc), based on stuff your competitors do etc etc.

Obsessing about conversion rate means a focus on the 20 – 40% of the traffic on your website that is “in the game” and solving for just that minority.

Obsessing about conversion rate means that you are implicitly ignoring major parts of the traffic that you should be creating optimal customer experiences for (pages, content, whatever).

Obsessing about conversion rate, perhaps most importantly, means you solve for the short term, the now, and just Submit Order,  at the cost of solving for long term metrics like creating Net Promoters.

(And I say that knowing that perhaps for your site 0.001% improvement in conversion is ten million dollars.)

So if you take me advice and mount this coup in your company and dislodge conversion rate from the throne which metric should take its place?

Let me submit one powerful metric for the coronation: Task Completion Rate by Primary Purpose.

Figure out what the the core reasons why people come to your website (the question to ask is: Why are you visiting the website today?) and then figure out if for customers in each of those segments of primary purpose if they are able to complete their task, whatever the task the customer came for (the question to ask is: Where you able to complete your task today?). Do a little cross tab and you’re in business.

The answer to Primary Purpose question will look something like this: Research Products/Services, Purchase Products/Services, Look for Company Information, Register the Products I have already Purchased, Looking for Support, etc etc.

The answer to the Task Completion Rate question will be Yes or No.

If you can do this you suddenly are 1) massively aware of why people come to your website 2) just a few of them are there to buy, and how many 3) where is your website failing you.

Now you know what you need to do to improve your website experience for your major segments of customers to increase their task completion rate. It is impossible that higher conversion rate will not follow.

But as you can see the behavior you will drive in your company will be fundamentally different under this approach. 

Let me close with a humble/arrogant (your choice) point of view by applying the recommendation to a real world practice.

I am a fan of Persuasion Architecture by Future Now Inc. I have had the privilege of hearing Bryan and Jeffery Eisenberg and John Quatro-von Tivadar speak, they are awesome (Future Now blog). I like the overall beauty of Persuasion Architecture, especially the Uncovery process.

Stated simply the overall approach is: Understand customers really well, create personas for customer segments and then use patented software from Future Now to redesign website to solve for those personas needs/wants leading to sales/conversion. 

Based on the logic of this post IMHO the overpowering focus on Conversion (getting visitors to buy) and doing all the work to get a site visitors to buy limits the massive impact that the methodology can have.

Rather than being “Persuasion to Conversion by solving for customer needs” of just buyers there would be a exponentially powerful outcome for Future Now and their clients if it was “Persuasion to perfect Task Completion Rate by solving for all customer segments”. Everything about the architecture stays the same. Except the final goal.

It is much harder to solve for multiple goals, to move from solving for conversion rate / buying to solving for task completion for your major personas and the purposes they come to the site for. But doing that will provide long term sustained competitive advantage that would be hard to replicate by your competitors. (Besides my mom says that nothing worth valuing in life comes easy. : ) )

In the end, when the dust settles, it is about creating satisfied site visitors no matter why they come to your website and secondarily it is about short term revenue.

What do you think? Too much la – la  – la? Anything I am missing in my thinking? Are you convinced and will you mount a coup in your company to dislodge conversion rate? Would you like to give me a hug? Please share your feedback via comments.

[Like this post? For more posts like this please click here.]

PS: This post is dedicated to my dear dear friend Blaire Hansen, for reasons she and I know :^).

We have already beaten this horse dead, several times over: Web analysis is hard, there is too much data, don’t know how much of it to trust.

Also yes your job as a web analyst is a very hard one, even if you overcome the data hurdles there is still the challenge of convincing people to listen to you and take action based on your analysis / insights.

Perhaps it is because we swim in “session_id’s” and  “shopper_id’s” and “evar’s” and “parameters” and “strings” that often the analysis that comes out of all that excellent work is very "dry". One of the reasons our decision makers don’t jump up and take action is they have a hard time relating to our analysis due to its “dryness”.

Hence Tip #4: Make data “connectable”. Quite simply this means that we should take as much of the “technical” / “dry” stuff out of our analysis, or should I say recommendations, and try to make it easy to understand from the perspective of non-analysts / business users.

This could mean using “pretty pictures” (slides or graphs or colorful tables or click density report screen shots), it could also mean using language and terms that “hide” some of the dryness and make our data easy to understand and to relate to.

There are tons of pretty picture examples (see my post on data visualization for one such example of "connecting" via pretty pictures). Hence I’ll illustrate this analytics tip with a language example.

I am sure you have done reports on “site engagement”. There are many ways to define site engagement but a definition frequently used is “number of pages viewed” on the website.

Perhaps a normal report for you is Site Engagement and Purchasers.

So you crank up your favorite analytics tool and produce this report (after applying the rules that: 1. usually absolute numbers are sub optimal 2. trends are magnificent and 3. segmentation is God’s gift to analysis):

This is a very insightful report and it is extremely easy to read and understand for most of us. As we look at it potential actions jump out at us right away.

But to a business, non-analytical, user perhaps they can pick up a clue or two from this but I would hypothesize that they might not get as much out of this as you might want them to.

Now let us try to make this “connectable”:

Ahhhh simple sprinkling of the english language and the table, with exact same data, “speaks” a lot more to any kinds of data consumer.

Now when our business decision makers see this data they are able to internalize it much more. They can relate to the english terms we have used much more then they could understand segments as defined by page views. The numbers are the same but this table will suggest action to our decision makers.

In real life when this was done here is what happened :

• The users wanted to know the definitions, this meant they actually bothered to learn what each segment was in a way that did not happen before.
• They looked beyond the site conversion rate (represented by the one-off-wonders, loyalists) and from the numbers truly understood the opportunity landscape that the first three sets of numbers represent.
• Since the terms used are very reflective of the persona way of thinking the terms became ingrained in the company culture (especially marketing). People started taking about Flirters and how to convert them into Browsers and meetings where held to specifically target Browsers and get them to become buyers and special programs were created for Loyalists (because they were so rare).

  You can’t underestimate the power of your customer segments becoming a part of decision making culture.

• Additional monthly analysis were created for each segment with someone assigned to “take care of the browsers and give them some love”, (translation: what action can we take?).

This is an extremely simple example of making data and analysis more “connectable” but hopefully it illustrates the power of stepping beyond our level of thinking and into the world of our decision makers and as a result doing something simple to have a big impact.

Choice of words and careful thinking can enliven the simplest or the most complex report helping us communicate our message more effectively, as well as in some cases atleast bring about some fundamental change in our organizations.

Agree? Does this make sense? Not buying this at all? Do you have examples where you did something different to make data more “connectable”? Please share your feedback and critique via comments.

Props : The original creative spark for this specific approach came from listening to a presentation by Tim Boughton of Holiday-Rentals.com & WVR Group at the eMetrics summit in London. I have evolved Tim’s suggestion a bit above. Suffice it to say the data and profiles above don’t remotely reflect that of Tim’s employer or mine, though if you find it useful please book a vacation at his employer’s website or buy some financial software from mine. : )

[Like this post? For more posts like this please click here.]