Feb 12010

Shopper Profiles and Conversion Rates Part I

There is data and there is the human behavior it seeks to describe.

Data driven marketing does not always demand an understanding of human motivation; drive by the numbers and the numbers will improve as a result. However, we often find that grasping at least some understanding of the underlying human behavior can help us better predict what strategies will and will not yield results.

Many explanations have been advanced to explain why measured conversion rates for general search terms are lower than those of more specific searches. No single explanation covers all the observed behavior, but one possible explanation had never occurred to me until recently.

Let's run through the standard explanations and assess their strengths and weaknesses.


A furniture retailer finds that conversion rates for the keyword "furniture" and "round end table" are very different and that the value of the traffic on the latter is 4 times that of the former though the volume is far less.


  1. The Buying Cycle: User's searching for "furniture" haven't decided what they want yet. These early stage shoppers will eventually return and purchase but those orders may be credited to other search keywords or other marketing programs entirely. The conversion differential is an illusion according to this argument.
    • Strength: It makes a great deal of logical sense, and serves as justification for larger ad budgets.
    • Weakness #1: In retail, the data just doesn't support the explanation. I'm told it's more prevalent in other verticals, but in retail this only explains a small portion of the effect.
    • Weakness #2: Do people really shop for "furniture"? "Honey, I'm going out to get some furniture, I'll be back in a couple hours!" Seems to me, most people who are shopping are looking for something in particular. I could be wrong, but I'd bet far more people shop for an individual piece -- the leather sofa -- than look for a whole roomful.
  2. Browsing not shopping: Users searching in general terms are mostly just kicking tires, whiling away the time virtually window shopping. People using more specific terms are serious buyers.
    • Strength: Compelling, and the conversion rates certainly support it.
    • Weakness: Wouldn't we see more buying cycle behavior if this were the case? If people in "buying mode" shopped using more specific terms, where window shoppers search more generally, we'd see a strong buying cycle in the data, and we'd see the vast majority of sales take place after specific searches. The fact that we see little of the buying cycle pattern and the majority of sales on more general "head" Keywords suggests this is not the root either.
  3. Failure to find what they seek: The more general the search, the more likely it is that the retailer's selection isn't comprehensive. With a SKU specific search retailers carrying that SKU have a 100% comprehensive selection for that particular model. No retailer has a 100% comprehensive selection of "furniture" so the chances of making a sale are reduced.
    • Strength: We see some evidence to support this. When a retailer carries a comprehensive selection of a particular manufacturer's brand, conversion rates on "manufacturer brand" are often just as strong as searches on "manufacturer brand + category" or "manufacturer brand + product".
    • Not really a "weakness" as much as a curiosity: Wouldn't this lead to frustration and therefore convince people to search with more specificity?

This last question, and another observation led me to think searching skill as an explanation.

My wife is a brilliant person: scientist, artist, and CEO. While she is my superior in almost every respect, she doesn't search as well as I do. The other day we tried to remember how Hamlet's "To be or not to be" speech continued. We turned to our respective laptops to race to see who could find it first. I won, and I always win these races. I searched for "Hamlet 'to be or not to be' soliloquy", she searched for "Shakespeare quotations". I found what I wanted on the first click. She went through a number of different links before she found a site with the full-text. We both found what we wanted but it took her 4 times as many clicks as it took me.

Searching skill seems to explain both why we don't see much of a buying cycle, and why we do see a huge volume of sales on general searches. We don't see successively specific searches very often because people who haven't learned to type in exactly what they want the first time likely haven't learned that the next time they search either. Yet, because this skill-level represents the majority of the population the sales volume is there.

In Part II of this series I'll lay out shopper profiles that might help explain some of what we see in the data.


7 Responses to "Shopper Profiles and Conversion Rates Part I"
Social comments and analytics for this post... This post was mentioned on Twitter by the9variables: Nice post via the RKG Blog on SEM: Buying Cycle or Skill? http://www.rimmkaufman.com/rkgblog/2010/02/01/buying-cycle-or-user-skill/...
Ophir says:
Hi George, Interesting post, very intereting. I find myself struggling with this issue day in day out and I mostly agreee with your insights, I just wanted to emphasize smth: When looking at heads vs. tails, I look at it this way: Hanging out in the mall - you go alne or with someone, look right and left, enjoy he window browsing, and if smth really catches your eye you would spend a few seconds/minutes more browsing it, you will most likely not buy anything specific. (Heads) BUT When you need a pair of shows, you'd browse only shoe stores, focus on those brands you prefer, try out a few and probably buy a pair or two (Tails) Question remains - why would you prefer that brand and those two pairs? How would you get to know they even exist? Answer: Heads.. Heads IMO are the demand and awarness drivers, and tails are simply those which convert. The funnel of course creates a situation where volume of the heads is so much higher than tails, that it may seem as one tye is enough - but kill one, and you may find youself barefoot... Ophir
Ophir, thank you for your marvelous comment. I agree with you. Brand building is an important element of marketing, and a very different animal from direct marketing. At the same time, where brand building is the crucial form of marketing for Nike, for a store that sells Nike among others brand building is a very different proposition. Yes, "I want to be like Mike" is powerful, demand generating stuff. Paying to have someone walk through your store and buy from a competitor's is a harder nickel to justify, I think. No question demand generating advertising matters, but as we've discussed, paid search is an expenses proposition for brand building.


Check out what others are saying...
[...] Shopper Profiles and Conversion Rates Part I, Rimm Kaufman [...]
[...] couple of months ago I made some observations about search behavior that I witnessed first hand. Namely, that people search differently from one another and that these [...]
[...] couple of months ago I made some observations about search behavior that I witnessed first hand. Namely, that people search differently from each other and that these [...]
[...] Part 1 of this post we argued that the lower conversion rates associated with general search terms may be [...]

Leave A Comment