How are Hummingbird and “not provided” affecting the Search Marketing industry? What are some insider tips and tricks? SEMPO recently chatted with Bill Hunt, President of Back Azimuth Consulting and co-author of Search Engine Marketing, Inc., about his views on Search Marketing and the latest industry changes.
1. What initially attracted you to work in the Search Marketing industry?
I got pulled into it indirectly by optimizing my own earthquake preparedness site to rank well in US and Japanese search engines. After the Kobe earthquake in 1995, we did significant business with Japan and that success was reported by a number of business magazines. The phones were ringing off the hook but not for the earthquake kits or consulting but from companies. Most wanted us to help them enter Japan using search engines. I sold the kit company and my wife Motoko and I focused on localization and optimization of sites for large companies like AT&T, HP, and Western Digital.
2. Tell us a little bit about your approach to gaining keyword insights.
My process is a multistep process to segment the words by logical categories and let the data point us in the right direction. Keywords from different sources can tell us a lot about our consumer.
The first and primary insight I want to know is how are we performing for our products and services. It is not sexy but performance for all products, services, categories and their buy cycle attributes. These don’t need to be researched or supported by any data – they are your primary keyword universe and should be non-negotiable.
Once we have the products and services data all set, what does the data tell us? Do any patterns emerge when you look at a cluster of words? Do they use one over another; is there a price relationship or a lack of knowledge about the product? Some interesting examples I have found are:
Las Vegas Hotels – what do they really mean when they use that phrase? We found that 83% of the search volume for variations of this phrase were related to “cheap or discount” hotel rooms. That while they searched using “Las Vegas Hotels” they really meant “cheap Las Vegas Hotels.” So if the #2 ranking page has a snippet of “The premier Resort on the Strip,” which clearly does not sound cheap, you are not going to get clicked.
Cloud Computing – along the same lines, we found 87% of the variations of this phrase were “what is” or “what are the benefits” related to cloud computing and all the pages ranking on the first page were educational pages that described what cloud computing was. Once my clients changed the content they jumped to the first page.
Some other sorts we want to look at are:
By revenue – what keywords make you money?
By margin – where do you make the most money?
Misspellings – how many ways do they misspell different words?
Currently I am finding one of the best sources of insight is site search data. These are phrases that are being used on your site showing specific interest in the product or service. For one of our retail clients, we identified over $4 million in untapped opportunity to upsell and cross sell site searchers who wanted to upgrade or add to their products but none of the content supported these phrases. Simply adding the ability to buy up recovered the cost of the effort in less then two weeks.
Another big analysis I think people should do is to look at their top 20 highest CPC paid search terms and see if they rank in the top 3 organic and which page is ranking. So far nearly 200 people have told me that less than half of their terms are ranking for their most expensive words. Google research shows that as much as 66% of the paid listings that are clicked don’t have a corresponding organic listing. This is pure gold if you can fix this not only from a cost reduction but from a brand and shelf space opportunity.
3. What kind of tools do you use and recommend to gain keyword insights or research? Does this differ based on the size of the company and, if so, how? What can be gained from drilling down through keyword data?
I honestly believe the process is the same no matter the size of the site or company. I think that some search teams get overwhelmed thinking about keyword data modeling for an enterprise company with millions of words.
The #1 tool is your brain. I don’t think enough people spend time thinking about their keywords, their audience, and, most importantly, what did they want when they searched and why did they use a specific phrase when they did it.
As far as commercial tools, this is the interesting thing about our industry – there are not any available. Sure all the tools have keywords in them and give you data and a few even have some basic classification but none are really enabling you to do anything with the words and data.
I am obviously partial to the keyword management/mining tool that I have spent the last 2 years developing – it pulls in all your keywords and associated data and does the segmentation I have already mentioned.
For those without a tool like mine, I think a simple database or pivot tables in Excel can work wonders. Once you have tagged the words into classifications, you can sort them any number of ways to find opportunity. I am finding most advanced search programs, like those at SAP and TripAdvisor, require their Search Team members to be proficient in SQL to be able to mine for opportunities.
There is a lot of data that can be gleaned from the data. I have hundreds of examples of opportunity that was mined.
4. Has Hummingbird impacted your process for recommending/prioritizing keywords or phrases?
It has not changed our process but has been helpful to get companies to better understand the need for Searcher Intent Modeling. This is especially true with any phrase related to “how to” do something. Hummingbird is trying to deduce the intent of the search and we have seen a number of pages drop for companies that are not specifically answering the “how to” or don’t specially match the intent of the searcher. Also we are pushing content type alignment – for one of our clients that has a lot of cocktail recipes, many of their web pages were replaced by their YouTube videos on making various drinks. We have made some changes so that the video and the page are ranking, increasing their SERP shelf space.
5. In your opinion, when should you trust your gut instincts and when should you trust data in keyword selection?
I think you have to always trust your gut first since it is a great gauge of intent. Search Marketers rely too heavily on tools to guide them rather than instinct.
6. How important is it to reevaluate/review keywords? When should this be done?
It is critical to do it at least quarterly, and I have a few clients that review all new keywords on a weekly basis. It really depends on the turnover of your words and how many new products you have.
7. How has “not provided” impacted your process of measuring keyword optimization success? What metrics do you recommend for evaluating keyword success?
It has had a pretty significant impact but we are starting to work around it. I am well known for my “Missed Opportunity Models,” which use the gap between search volume and organic visits for a specific keyword. These models have been used by many companies to justify their search programs and they are nearly impossible to do now.
Most of the companies I work with still use rank and organic search engine revenue as their key drivers. While we cannot tie it to a specific keyword, we can show that overall non-paid search revenue is increasing.
We just built into our tool the ability to leverage PLP’s, rank, and page revenue data as a proxy for organic performance. How this works is every Tier 1 and Tier 2 keyword we have assigned a “preferred landing page” – the page we believe has the best opportunity for conversion. We then look at organic attributed revenue month over month and rank of the assortment of words month over month. This allows us to show that an increase in ranking or the swap of a PLP over another page has had positive impact. We can see the same when a word or a set of words drop in rank.
We also look at traffic based on Google Webmaster Tool data which we pull in but it is only a fraction of actual data.
8. Where do you see the role of keyword research and optimization going in the future?
They are becoming even more important, especially since we have less data to support performance improvements with not provided. I have been advocating a Keyword Czar role for a number of years, and there are a few companies that are starting to adopt this at least as a part-time role. With the increase in social media conversational mining and the demands for content ideas, this is becoming more and more important.
It is a great tool for content creation efforts. We do a lot of modeling of words and opportunity and if you lay that out in a hierarchical flow it pretty much tells you exactly what content your are missing and how to better speak the language of your consumer.
How are Hummingbird and “not provided” affecting you? Where do you see the role of keyword research and optimization going in the future? Share your thoughts in the comments.
Opinions expressed in the article are those of the guest author and not necessarily SEMPO.