Sun.com Customer Experience and Stuff: Notes from the Sun.com customer experience team.

Have you ever been lured into a store with high hopes of making a great purchase, only to be turned away by long lines, unfriendly sales people, or a confusing store layout?  If so, you were a victim of a bad experience, and so was the store since they didn't get your money, and maybe lost you forever as a customer.

In marketing (and now commonly also in experience design) this situation is called a "conversion problem" -- the store failed to "convert" you into a customer even though they did get you to enter the store.

Conversion modeling has been treated as a distinct science in the direct marketing world for decades, where prospective customers are sliced and diced into "segments" and then conversion rates of different mailing campaigns are tracked to compare their effectiveness.

In the customer experience world, many of us build models of all of the common steps in an experience, and then look at the conversion rates between each step to understand where customers are getting stuck and bailing out. For instance, here is a picture of a simple modeling tool we developed at Sun, using an imaginary product from an imaginary company (see the end of this posting for a link to a StarOffice spreadsheet model):



How do you use such a tool to improve customer experiences?  The idea is to understand the steps that a user goes through to get to a goal, and to optimize the "conversion rate" at each step, with an eye also on the total conversion rate. The picture above uses web pages for illustration, but you can use conversion modeling for any experience that involves multiple steps.

For instance, in a physical store, you could define key milestone steps of a buying experience such as

1. Being attracted to the store
2. Entering the store
3. Assessing whether the store sells the kind of thing you're looking for (e.g. lipstick)
4. Identifying the specific product you want (maybe with helpful advice from a salesperson)
5. Buying other related items nearby (being "cross-sold" on make up or trendy lip glass kits)
6. Queing for checkup
7. Paying
8. Establishing on ongoing relationship (e.g. email sign-up)
9. Leaving the store

If you think about it, that's a lot of steps, and a lot of places for things to go wrong. For instance, if only 50% of lipstick shoppers who walk in the store can tell whether it sells lipstick (the first step), half of them will leave. Right off the bat.  You've lost half your customers already! And it compounds. If there are eight steps in a hypothetical shopping process and even if you keep 80% at every stage, your total conversion rate is 80% x 80% x 80% x 80% x 80% x 80% x 80% x 80% = 16.8%. That is, of 100 potential buyers who come in, you could expect to retain only about 17 to the end.  And may of those twho abandoned you probably did so because of a confusing or frutsrating experience.

We model experiences in a similar way in the online world. Once we understand wheree people are abandoning the web site, , we can use other techniques such as usability tests or surveys  of traffic analysis to funderstand the problem and fix it.

Conversion modeling doesn't have to be about money. You can model any experience that involves free stuff like downloads or newsletter signups. The idea is to understand bottlenecks and get them out of the way so that your customers have a more streamlined experience.

It wasn't too long ago that web folks and designers shunned techniques from direct marketing, considering them old fashioned and quaint. But now most online companies are obsessed with conversion, because they know that good conversion rates translate to the bottom line.

I've posted a template you can use  to model conversions. It's in StarOffice 8 format.  Let me know how it works for you.

P.S. Though the conversion funnel idea is simple, there is certainly nuance. For instance, anyone who's analyzed web site traffic knows that users arrive and travel on various tributary paths, so there is no perfect funnel in many cases. And, sometimes, it's OK to have non-target users or customers abandon pages early so that you route them out of the process -- the customers you care about in such cases.

Technorati Tags: Design Web-Design CustomerExperience

Tunes: 73: Willow Weep for Me: Duke Ellington