Frustrated at trying to find a list of the top 10 all-time best selling women’s fragrances on Google, it dawned on me that increasingly we are looking for specific answers to specific questions rather than a trove of documents on a particular topic. It forced me to think about the changing nature of search and how quickly search has become an almost automatic reflex.
And while we often focus on the accuracy of using search to target customers, we don't really spend much time analyzing the accuracy of search results. After all Google has set our baseline expectations with "good enough" search that's fast and free. So who are we to question how they get results? In fact, most of us are so focused on the instant gratification that we don't really know or care how it came about. But as search becomes completely integrated into our thinking process differences in the way data is found and the immediate relevance of the data presented will become more important to us. It seems like a natural developmental path; that as we use and understand the tool we want more from it and want it to be even better.
Maybe this is the implication of new research from Hitwise that indicates that longer search queries are growing fastest. Search queries of 6 words or more increased between 8 and 20 percent over the last year while one and two word searches posted slight declines. To some extent this has to reflect increased understanding of how to search. It also suggests that people are using more words in an attempt to refine or better define the query in the hope of getting more accurate results. Its a yearning for the answer; not for the source documents. As we get used to the idea that vast amounts of knowledge are available to use, we want our clicks to zero-in and present the answer to us. We want search engines to go the last mile for us.
It’s not just that we are greedy or ungrateful. This idea of fishing out the answer has intrigued computer scientists forever. It has led Kosmix to build complex taxonomies to sort through and cluster data. It has driven the creation of high powered math to mine and organize relational databases that underlie Freebase, Powerset and Hakia. And it promises to leverage the promise of semantic search in the forthcoming Wolfram Alpha tool which is being hyped as the super-searchable digital library of Alexandria on steroids. Yet in spite of all this promise, search still needs to discern instantly what millions of people using the same vocabulary and syntax actually mean and actually want.
To that end, all this philosophizing suggests a few concrete things marketers should be doing in designing SEO or pay-per-click (PPC) campaigns.
Compile Answers. Look at your business and develop a list of the most frequently sought answers. Build the answers into your sites, press releases, videos, etc in ways that can expedite search. In many cases these will be obvious and probably already there in some form but not necessarily served up succinctly. Consider simplifying and expanding your FAQs.
Test Longer Search Phrases. Analyze your traffic and mirror customer and prospect search behavior. Experiment with longer, alternative ways to describe or explain your goods or services. Test different words and word sequences to see if they return more relevant results. I suspect that longer phrases might also be cheaper than shorter phrases especially in categories dominated by generic or name brand terms, at least until everyone catches on.
Use Terms that Impact Sales. For most things there are predictable sales stages and inflection points. If you think carefully about these stages/points and buy or embed the phrases that reflect them you can more directly influence the awareness-consideration-con version sequence. Maybe more descriptive, more accurate or more complex phrases can bring prospects through the pipeline faster.