Chocolaty landing page segmentation

Directed user segmentation is a mouthful. Once online marketers understand what it is and how to do it, they wonder how they ever lived without it. And it actually becomes quite simple. But it’s hard to explain without droning on in what sounds like very technical and boring terms. I’d like to shed all that babble and focus on chocolate.

Chocolaty-rich landing page parable

Let’s pretend you’ve got a new chocolate to introduce. And it’s like the best chocolate on earth. But nobody knows that yet, because nobody’s tried it. You know chocolate lovers will respond and evangelize your chocolaty goodness — if you can just get them to try it. You need to put free samples into the mouths of these people.

You decide on search engine marketing to generate free sample requests — and build your seed database of chocoholics. You know these wired connoisseurs are jaded — they’ve tried every chocolate that claims the same things you’re claiming. You need an edge.

You realize that all chocolate aficionados are not alike. In fact, there are several distinct sects within the cult of chocolatopia and their tastes are very different. There are the bakers and the gnawers, the darks, the milks, and even the whites — but they’re not in your target.

Your search ads reach out for passionate chocolate lovers. That wider net gives you the traffic volume you need to move your cocoa-infused madness closer to global domination.

Click to use this example conversion path. It has two levels of directed user segmentation. WARNING: It’ll make you hungry.You drive hundreds of thousands of clicks to this conversion path. It segments your chocolate lovers into like-minded groups that you can persuade to partake of your semisweet chips, dips and bars. You speak to each group’s tastes and needs in language that makes them hungry for more. You send them chocolate. They love it. They spread the word. You rule the chocolate world.

All because of directed user segmentation.

Comments

[RJ]

Good example of your post click philosophy. My only thought is that how do you track and account for those that like both segments? They follow one then back track to follow the other? I guess in the end you still hopefully get the registration but it skews the data from the segmentation. Your thoughts?

[ion]

We record the initial segmentation choice as 'primary' and subsequent tags as 'secondary'. A relatively small minority of users will back track and try another branch to make sure they're not missing out on something, but typically they end up converting on the original branch as that was truly what interested them most. We have seen no correlation between back-tracking and abandoning. To the contrary, back trackers seem to be more engaged and thus more likely to ultimately convert.

I can say that better segmentation design seems to result in fewer back trackers. And a more credible brand and message increases trust which also appears to decrease the likelihood of back tracking.

Thanks for reading.

[R4 revolution]

Nice, keep posting such stuff in the future as well. We will be looking at you. Thanks

[Emily]

How would you segment in this case? I am a company that sells software, and a person does a Google search for "Microsoft Office". Well, there are multiple editions we carry: XP, 2003, 2007, Mac. Within those editions there are multiple versions: Home & Student, Standard, Small Business, Professional, Ultimate, etc.

Gotta love Microsoft for this. But to be honest this is the case with many of the software vendors we sell. How do we segment these landing pages with so many options?

[everyman]

Emily's question illustrates my point I made on another post. For cases where the choices are few, obviously self-selection makes sense. Once you get past a certain level of complexity, intelligent automation will help. Think of a store on Etsy vs. Amazon.com - the closer you get to having a sophisticated site, manual winnowing is not going to cut it.

[ion]

I think Emily's challenge is one of information architecture more than segmentation. Could she segment respondents into buckets to match them up with the version best suited to them? Yes, but would the experience of that segmentation fit well with searcher intent? Unlikely.

Office is a ubiquitous product. Generally speaking, respondents will already know what Office is and will likely already be predisposed to make a purchase if: they're confident they're getting the best price and they feel assured that they're buying the version best suited to their needs. Since those criteria are so black and white, I think the challenge is bringing clarity to Microsoft's complexity.

Matching needs with versions is less about segmentation and more about configuration. You're asking questions that have very black and white answers. Your operating system is your operating system — you're not going to get one over the other because of who you are or what you're going to use Office to do. Beyond the OS, you have licensing (and in some cases feature differences that go along with those licensing alternatives (I believe)). Again, it's very black and white. For the license, I would test a simple, well designed table of differences against a set of descriptions and see which works better.

Thanks for reading.

[ion]

Segmentation works well for warming pages that precede complex, subscription, or high-ticket items. It is far less effective (and in fact gets in the way) in the context of a vast array of simple, or well understood products.

If a searcher is looking for a widget, give them the widget. But if a searcher is trying to find a fit, help them. Segmenting to discover needs and match those needs to solutions is highly effective at increasing e-commerce conversion rates (and average sale values too).

Read our American Greetings case study for a good example of lowering CPA using segment-specific warming pages on a mass, lower-ticket scale.