The Evolution of Personalized Search Results and the Search Marketer

[L]

As we enter the age of an inevitable personalization of search results, I, like others, cant help but wonder what the future of the SEO/SEM professional may be. Will we all mesh into the reporting world as part search optimizer, part information analyst? Sure, many of us are on that road already, but if your heart is in SEO, and your daily tasks are getting front and center rankings for your company or your clients, you’re an SEO. But what will you be a year from now?

As I scrutinize our rankings in GYM (Google, Yahoo & MSN) every day and create reports and analysis of traffic and search terms and ranking relationships, a little voice keeps sneaking into the back of conscience saying “Hey! Personalized rankings are going to sweep the nation and no two people will have the same search results, so how will you know where you rank then, sport, huh?”

Sure, there are those who preach not to be so obsessed with rankings, it’s all about conversion. There are countless threads on this topic. I absolutely agree, conversion it is. But conversion is just a part of the whole, a later step in the process. As any marketer knows, you need to get in front of your target markets before they can possibly convert. How effective is measuring only conversion if you don’t know how well you’re even reaching your audience? And how do you know if you’re effectively reaching your target audience in search? You search! You see if you are on page 1 or page 100, and you can make a pretty well-educated guess at what kind of traffic you might receive from either of those positions. Hence, knowing your ranking in SERPS is key.

So when personalized search comes into play, and each searcher will see their own unique set of search results, this leaves us with being able to tell what search term a person used to come into your site, and from what search engine, and whether or not that search term converted, but do we know anything about that person’s search experience before entering? Did they receive some sort of search results manipulated towards their preferences? If so, can we segment this person into some sort of user type? How will we know what their search experience was? As far as I can see it, there is no way to know. BUT I think with the help of the search engines, we may be able to gather some of that information for our own reporting and smarter marketing.

Don’t get me wrong – I’m all for personalized search from the user’s standpoint. It makes perfect sense to give the user a results set that is likely to be what they’re looking for. To steal an example from Personalizing Search via Automated Analysis of Interests and Activities, if a person does a search for IR they might be looking for infrared light, information retrieval, or maybe Ingersoll-Rand. By using other information to understand what the user might be interested in, the search engine can give the user a more relevant set of search results. It can make a better guess whether that person was looking for infrared light or Ingersoll-Rand stocks.

The question is, when I’m looking at my site logs, how do I know if this set of search results was manipulated for this person, and how? Other than having some vague demographic knowledge about the typical Google user, how can I determine what type of person this is? If I see that the visitor entered my infrared light web page from MSN with the search term IR, what assumptions can I make as to how well I did as an SEO to rank for that term? I can guess that if they converted, I must be ranking well for my target market, assuming they are receiving these personalized search results. If I see that people are coming from MSN with the term IR, and leaving immediately, I may be appearing in search results for the wrong market. Which leaves me with the task of positioning myself better within my own industry, and probably adding more relevant supplemental terms to my page that have to do with infrared light. Which really isn’t much different than what I do now. Will conversion be the only way to tell whether or not you’re successfully appearing in front of your target market now? How will we know whether or not we are more successful than our competitors at reaching our target markets? What if the Search engines could pass some information on what rank you were in that personas search result? What if they could pass along a sort of segmentation, industry, user type?

[L]

Now let me break it down even further. Many personalized search engines are still testing out what works and what doesn’t. Lets take two examples from Google: Google personalized search and Google’s segmentation of search results into sections on a page.

The first currently requires the user to login, and once they have done so, they will receive their own personalized home page. This page currently allows the user to add their own feeds to the page, so they can receive whatever information they want to watch every day; weather, news, industry top stories, etc. All right there on your Google search page. On top of that, your search history is watched and possibly now or soon, everything you bookmark, the emails and documents you read and write, every site you go to and return to and how long you’ve been there and how often you go back, etc. Now when you search for IR, Google has some insight into what you really might be looking for. Hence if your into spectroscopy and you were just reading about Simultaneous quantitative analysis of four hydrocarbon gases in a mixture at ppb-level concentrations using EP-IR spectrometry, then Google may be more likely to serve up some infrared results for your “IR” search rather than information retrieval or Ingersoll-Rand.

The second Google SERP manipulation which may or may not still be in experimental mode, is to segment a set of search results into different types of results on one page. For example, as of this writing if I search for IR (I’m not signed into personalized search for this test, but interestingly enough, I get the same 1st page results right now whether I am signed in or not), the first five result titles are:
International Rectifier Home Page - Flash Detection
Irna
WebSpectra - Problems in NMR and IR Spectroscopy
اخبار رياست جمهوري
IR Spectroscopy Tutorial

Then below those results there is a horizontal line divider and Google says See results for: inland revenue with these results listed under that:
Inland Revenue - Te Tari Taake
Internal Revenue Service
HM Revenue & Customs: Home Page

And then another horizontal line divider with these two interesting results:
Information Research: an international electronic journal
Ingersoll-Rand Co Ltd

And that is my first page SERP. Google is giving suggestions for a broad, general term that could take on multiple meanings, depending on what the searcher actually had in mind.

Now how do I explain to my colleagues in n months if/when this personalized search craze becomes the norm, that this rank reporting that we are putting so much time and effort into, is obsolete? That our # 6 ranking in Google for the term IR doesn’t really mean that we’re number 6 since its cut off by a horizontal rule after the fifth result and brings in other sites? OR, maybe we are still ranked number 6 if the API actually tells us that the number 6 result is inland revenue, and that’s my site – even though we’re listed in the second segmentation under the horizontal divider of the page, we’re technically still #6… When I’m looking at my logs and analyzing my reports like a good information analyst slash SEO, how do I know what the user experience was before they came to me? Can we skew the tool to know more about our searchers, just like the search engine will know more about our searchers? To do that, we’d need more information passed from the search engine. I’m not completely sure what that information would be just yet (at least rank?), but I do know that I want to understand more about what my visitors user experience was before coming into the site.

In any other marketing avenue, understanding this is key. For example, if I’m running a print ad on an elevator overhead at LAX, before I run that ad I know how may people take that elevator every day, what planes they came off of, what time of day they go through, what type of people they are (business travelers, leisure travelers), how much money they make, what type of homes they have, where they go every year, and more. Now that search is becoming a more sophisticated marketing arena, how do we get this information?

When a person is logged into Google, Google has the chance to capture demographic information about that person. It is not uncommon to have to enter your address, age, sex, or other criteria when registering for anything. Google also has, however slighted it may be, IP information from wherever that logged in Googler is Googling from. Here’s some demographics. This type of information may soon to be available from some search engines for a price. Probably a price only big companies with big budgets will be able to afford. So what about in our search logs? I’d love to be able to find a solution that will allow any of us IR (information retrieval) geeks to have some sort of insight into how we came across to what market, before they entered our site and either converted or didn’t.

I’d love to be able to evolve my reporting along with the search engines evolution, instead of ditching it in six months to a year. If Yahoo charges buku bucks for demographic information and Google passes it on through click information, hmmmm…. Won’t we just love Google even more? If MSN does it…

The evolution of search marketing towards more traditional marketing may also help to weed out those self-proclaimed SEO gurus who read an article or two on search marketing, get clients, and then get them bumped from search results. To be in the game, and to be successful, you’ll need to be a true marketer.

And in the end, I hope to see some sort of evolution of our information on the front end of the reporting, when searcher comes into my site from a search engine, what information can we gather at this point and how? Rather than losing information (ranking) from search engines as we move towards personalized results, can we get more information and move towards smarter, more informed marketing like in traditional marketing areas?