In Part 1 of this post we argued that the lower conversion rates associated with general search terms may be best explained as a product of poor searching skills. Because people type in searches that are more general than what they actually seek, there is a greater chance that any given advertiser doesn’t have what they wanted.

The desire to classify is hard to resist, so I’ll throw out 3 different Search Shopper Profiles for consideration.

1. Traditional Tracy: Tracy models her search shopping behavior on how she’s traditionally shopped. A trained brick and mortar shopper she visits stores one after another until she finds what she wants.

   Online, Tracy doesn’t search for a “silk blouse.” When she’s looking for a blouse she will visit: Anne Tyler, then Talbots, then Lands’ End, etc until she finds what she wants. Conversion rates on brand searches are huge because she’s already narrowed the list of stores based on past experience, offline marketing, word of mouth, etc.

   She is not interested in sifting through search results to find a retailer she likes. She’s a brand loyalist; not exclusive to any brand, but less interested in taking a chance on an unfamiliar store than others might be.

2. Yellow Page Paul: Paul knows that when he needs to shop the way to find a store is in the Yellow Pages. When Paul wants a new Barcalounger he looks for the type of store that would carry one: a furniture store. The directory wouldn’t have a listing for Barcalounger, so Paul’s been trained not to look for one.

   Paul didn’t go to the white pages to look for a specific store, he’s not a loyalist. He’s happy to buy from any retailer that has what he wants at a reasonable price.

3. Modern Mary: Modern Mary knows search engines and uses them well. She recognizes the efficiency of specificity and is far more likely to use three, four and five words to refine the SERP and get to exactly what she wants rapidly. More comfortable with computers, she likely reads the ad copy before clicking on links realizing that that can save time.

   Mary’s proficiency with search may also make her more likely to consult comparison shopping engines and coupon sites as a way of finding the best deal.

If there is any validity to these archetypes we should see some telling patterns in the data.

• We’d expect to see Tracy-like brand buyers have a disproportionate likelihood of having other brand search records on her browser, and we’d expect to find lower incidence of very specific search queries.
• We’d expect Paul to have other broad search phrases mixed with some navigational searches but a lower propensity for very specific search queries as well.
• Mary might reveal herself as a candidate for the buying cycle. She’s comfortable with search, and impatient with unproductive visits. She may be more likely to refine her search when the initial SERP doesn’t have what she wants. Indeed, it wouldn’t surprise me to find that the click to order interval for Mary is actually longer than for Paul; the reverse of what the standard buying cycle argument would suggest.

We’re blessed at the Rimm-Kaufman Group with a large slug of data from many great retailers, and a burning curiosity to find answers to these types of questions.

Sadly, the exigencies of running a business require me on occasion to do some productive work, so for the moment I throw these theories out to the wind with no empirical foundation to support them. If we can dig up sufficient evidence to back these crazy claims there will be a Part III of this post; if not, we’ll just forget this little incident took place :-)

George