Smargeting - smarter targeting for more effective marketing

Targeting. It's quite possibly the most important factor in any advertising or marketing campaign, and for good reason: it works. But to reap results, you've got to get it right. And that, of course, is easier said than done.

In theory, effective targeting shouldn't be quite such a challenge. After all, its premise is fairly straightforward: find the right product, find the right user and introduce the two.

But as most publishers know, getting those clicks, leads or purchases can be tough. After all, it's estimated that just a couple of people in a hundred will even think about going near your ads. In which case, you'll need to do more to persuade them.

Test your smargeting options

In order to smarget, we'll be thinking about ads from two points of view: the way they look and the way they earn. There's plenty that's already been written about ways to enhance ad performance, and any legitimate method that optimizes revenue clearly has a place in smargeting. Any info you already have regarding ad positioning, blending and the like is a valuable resource, particularly if you feel it works.

But we're about to look at a smargeting concept that may not be so obvious.

The received wisdom is that 'blending' ads makes them more successful. In other words, if an ad's colors and look closely match its environment, users are more likely to click it. The reasoning behind this is so-called ad blindness, the idea that if ads stands out too much, most users will simply ignore them.

But what if, for some others, ads are something they want to notice? A part of a website they actually enjoy?

It's certainly true that many sites aimed at a particular demographic - younger, outgoing - use quite the opposite technique. MySpace, to take just one example, clearly differentiates its highly colorful, often animated ads from a very plain background.

Such 'psycho-blending' is a clear attempt to smarget a user's mindset and personality. After all, if ad campaigns in the offline world are created for particular audiences, why shouldn't online strategies at least partly follow suit?

On the other hand, as users become much more used to the tactics of 'classic' blending, a general drop in clickthrough rates seems to indicate that they're also learning to ignore anything 'camouflaged' on a page. (I'll explain exactly why I think this is happening and how to off-set its effects later in the article).

So which way to turn? Although there's constant debate about which system works best, you'll need to discover which gets results for you. And it's not just a question of tweaking ad colors because if you're smargeting, you'll want to investigate the idea of psycho-blending, too. Are some ad formats more suited to your users' overall personality than others?

The first step is to make sure you know exactly which different formats are available. There are many affiliate ad programs, and each have their own distinctive looks or functionality. Peel-back ads; classic text ads; text ads featuring images; video; you name it, you're likely to find it. right here in clickspiration. Sign up to a selection of programs offering ads you feel might work - or even some you think might not.

You'll now need to test these different ad types to compare performance on your site. This, admittedly, is far from an exact science, but harnessing the power of keywords will help balance results. And despite flaws, this simple test method should tell you far more about smargeting your users than you already know.

Work with a page that has no other advertising on it, and decide on a specific keyword. Ensure it's a fairly standard one so that ads are readily available across all formats; choose one, too, that you know attracts users - we want to test ad types, not the keyword itself.

As an example, I'm going to select 'camera', and will imagine that we're going to test inline text, a standard text ad, a mini-mall and finally a banner.

To work with the inline text format, you'll need an ad scheme that lets you isolate keywords as well as the number of ads you want appearing in the text - in our case, it'll be just the one.

To ensure that 'camera' is the only operative word on the page, simply add the 'sectioning' tags to your source code as specified by the affiliate program.

Publish your page and leave it online long enough to get a handle on the stats. Essentially, you're looking at clickthrough rate - page impressions divided by number of clicks - and should be able to track this data in your affiliate admin area or through your own stats software.

You'll want to ensure you allow a good number of impressions for each ad type tested - thousands preferably, but you can settle for less if your traffic isn't that high.

Once you think you've a good idea how the ad is performing, try the standard text ad. Place it inside your text roughly in the same position as the previous example, and make sure the size isn't hugely disproportionate to the inline link you've just used. Again, you may need to section keywords so that 'camera' is active.

Publish the page and check your stats. Once you've reached the set amount of impressions, repeat the process using the mini-mall format with 'camera' as a product. Finally, test a banner (again, you may need to isolate keywords or locate a specific camera ad from a company such as Commission Junction.)

Did you learn anything from these tests? Was there a definite preference amongst your visitors? And more importantly, how did clickthrough rates compare to the format you've been using so far?

If any are noticeably higher than your current format, you've learnt something valuable about your users. Either that overall, they prefer a different format to the one you've been using; or, that changing or mixing ad types provokes real interest.

Either way, you can use such results to push your smargeting further.