Search Engine Strategies in San Jose is coming up in about two weeks time. Representing Unica’s web analytics solutions, I will have the pleasure of joining Omniture, Webtrends, and Google Analytics in staffing a panel discussion on measuring success. This time, moderator Allan Dick from VintageTub has asked each of the panelists to prepare a short presentation on a specific topic of interest. The topic I drew is how to measure the impact of Search on the advertiser’s brand equity
Oh boy. Tough question. Web analytics is not ideal for this purpose. But as I dove in, it turned out to be a very interesting one! And it is somewhat reminiscent of many other kinds of multichannel marketing measurements, actually.
The main reason we do SEO and pay-per-click advertising is, of course … to get clickthroughs and conversions. Doh! However, come to think of it, we do not actually want everyone to click. For example, we want buyers who search for a product XYZ that we carry to click. But we would rather not pay for non-buyers to click on our paid search ads, really. We just would very much like for them to remember that they saw our brand name. Next time somebody asks them where they can buy this product XYZ, we would hope they might remember the impression that our ad left and recall our brand.
When people search for “product XYZ” and clicktrough to a web site, that is really easy to measure with web analytics. However, what measurable effect would this kind of word of mouth referral have? Well, as a result of the referral the recipient would either type in www.YourBrand.com directly which is something that we can measure and trend. Or, they might not remember exactly and go to Google type in something like “YourBrand product XYZ ABCD”. Your listing would appear and hopefully get clicked. That is another thing we can measure with web analytics. (In the latter case, you can also use tools like Google Trends just to see the search volume for your branded keywords)
So web analytics can give us an idea about our brand name recognition/recollection. But now the hard part. How are we going to prove that the reason why there is an increased search for our brand name keywords is because of preceding Search marketing impressions on product category keywords (and/or word of mouth marketing? After all, we are constantly in a gazillion marketing initiatives trying to bring our name in front of everybody at tradeshows, PR, direct mail, cold calling, broadcast and outdoor advertiisng.
I am not aware that there is a good answer for this question unfortunately. For display ads, there is at least the concept of a viewthrough. You see an ad banner somewhere, unknown to you, Doubleclick or Atlas or Trueffect set a cookie. Then two weeks later when you visit that advertiser’s web site your cookie is reconciled with the ad impression and your visit counted as a viewthrough. Well, the same just does not exist for search marketing, as far as I know. You can’t set a cookie on your search result listing that shows up on Google, Yahoo, etc. Even if one could set a cookie we would still fall down in the word-of-mouth part. It is one person who sees the ad, passes the referral, and somebody else who visits the site. Oh boy.
There is one trick that you can try however. You can use Hitwise, Google Trends, and similar tools from Yahoo! and MSN to plot the trend of search on your category keywords, say “web analytics solutions”.
You know that everybody who searched for your high ranked search terms probably saw your ad or organic listing. Now, if you plot the trend of direct and brand keyword visitors and find that trend to follow the first trend, then you can try an argue that there is a correlation. Ok, tough. But I am not aware that web analytics can do any better to prove the impact of search ads on your brand when there is no clickthrough nor view through. What we are looking for would ideally be something similar to the following:
By the way, brand equity is of course not just about recollection. It is also about positivity (thumb up or down?), functionality (e.g. is a Ford Escape an awesome hybrid SUV or an awesome offroad vehicle?), and actionability (i.e. how much do these customers ultimately buy over life time?). Web analytics can contribute measurements for some of these as well. But this blog post is already way too long to dive into that! So next time.
