Published on February 2nd, 2013 | by Derek Slater0
SEO for B2B: A miniature case study
This article was originally published March 2010 on emediavitals.com.
I believe it was Eleanor Roosevelt who said “Nothing lasts forever, and that goes double for a homegrown ColdFusion CMS.”
Okay, Ms. Roosevelt did not have the foresight to say this. But I have the hindsight.
By late 2007 my website‘s publishing platform was overdue for a refresh. It had been cobbled together on the cheap earlier in the decade, and the cobblers were no longer around. New content was a challenge; reworking old content was out of the question. So we had a nice library of award-winning magazine content crammed in a couple of different databases – and some flat files! – all with clever magazine-style headlines and very few internal links.
With great anticipation we awaited our planned re-launch on Interwoven’s TeamSite platform in early 2008. It was, as I said at the time, like trading in a Yugo and being given the keys to a Maserati. We knew we would sideswipe a few fire hydrants along the way, but we were darn sure going to put the gas pedal on the floor!
It was during the year leading up to this conversion that I started studying SEO intensely, fueling up with information from Search Engine Journal and so on, as detailed previously on eMedia Vitals. So there were two critical SEO-related elements in our plans for this new platform:
- One, a high-speed spin through the archives to relabel all the major features.
- Two, a rudimentary keyword matrix to map our internal linking and some of our new content.
It’s useful to look back and see where these plans worked and where they would have benefited from a tuneup.
Retitling for SEO
The archived content was ported into TeamSite a few weeks ahead of the go-live date.
We took back copies of the magazine, divvied up the years, and started tearing through the TeamSite database and re-titling all the major articles. Clarity over cleverness: That was our mantra. With six years’ worth of content to address, this involved a significant investment of time – although much of the work was done late at night, as noted in this faintly ill-tempered post on my personal blog. Frankly, this is dull and repetitive work unless you really love your site.
We chose our new headlines in a non-scientific manner. We didn’t use Google Trends or look at on-site search data; we simply tried to reflect accurately what the article was about.
This SEO effort went a bit further. Sometimes an archived discursive feature lead would push the first use of the keyword down three or five paragraphs. Often in those cases I inserted a new first sentence using the keyword, for the benefit of both the engines and any reader wondering why this article allegedly about security instead started with an appreciation of a Wassily Kandinsky painting. Real example. Great article, but not optimal for the online paradigm.
This work was done in a hurried manner, in part simply because we had to devote most of our energy to new content, but also because we had to do other maintenance and cleanup work at the same time. Many of the articles had formatting that didn’t translate into the new platform, so there were odd ASCII characters, missing line breaks, broken links to clean up. So even with a high work rate, this project continued for many months.
A person on the sales operations team was involved in a similar round-the-clock manual retagging/reworking exercise on ad-related materials. In a distracted and haggard state we famously almost bashed our cars together in the parking lot a few days before relaunch. (Isn’t site work such a wonderful bonding experience!)
The SEO Matrix
I developed a keyword matrix. Again, it wasn’t prioritized according to Google Trends or Adwords, and I don’t believe I even used Google to identify related search terms. I simply tried to systematically write down topics and subtopics of importance to our print readers. An SEO consultant would have laughed, but it was at least a rudimentary attempt at systematic thinking about all the topics of value to our readers.
I used this matrix in two ways.
First, for each of my major keywords, I selected the article that seemed most relevant and authoritative – again according to my own estimation, although as the months went on I was able to start using our on-site Goole search to see what the engine thought was our best candidate. Next to each keyword on the matrix, I noted the URL for this target article, then found three or four other articles on the site that included the keywords and linked those keywords to the main target article. As explained previously, that internal linking serves your readers, earns page views, and tells search engines what you think is important.
Second, for some keywords there wasn’t a great target article. So I commissioned several articles designed as topical primers for these keywords.
This tactic may call to mind the hotly debated Demand Media approach (see Sean Blanda’s excellent look behind the curtain, and also this slightly older Wired article). However, you can take this approach at whatever quality level you choose. I spent good money and got some GREAT articles this way, e.g. Incident Response Basics and Red Team Blue Team Simulations. Both of those pieces wound up becoming magazine cover stories as well as online stalwarts.
Was This SEO Rework Worth the Effort?
I would answer this question with a resounding “yes.” Whether I can persuade you to agree, I am not sure.
For context, “we” in this article means a team of three editor/writers, including me. I don’t have a record of the person-hours spent rewriting headlines. This work was done sporadically during office hours during the month or six weeks surrounding relaunch, but I spent many more hours on this work burning the midnight oil, continuing for at least a year past the switchover. Certainly hundreds of hours in total.
I can’t put much stock in before-and-after total search traffic to the site. Too many variables in play. There were technical changes in addition to content changes. This kind of platform shift can be quite disruptive to your relationship with search engines. Concrete example: Here’s an article about protecting intellectual property that did NOT change in the system switchover. Same headline and everything else. This particular article consistently got web traffic prior to the conversion, fell totally off the map the minute the new platform went live, and then bubbled back up in Omniture more than a year later. Currently it ranks #2 when I Google “intellectual property protection”. I haven’t touched it. In fact, now that I look at it again, it needs some cleanup.
I can point to specific articles that now rank well for a search term that was not in the old headline:
- “Unwelcome Diversions” became Unwelcome (Product) Diversions. Current rank for search “product diversion” is #6.
- “Success Factors” became Strong Authentication for Online Banking: Success Factors. Rank for ”online banking authentication” is #1.
- “Now Hear This!” became Security Awareness Programs: Now Hear This. Rank for “security awareness program” is #9.
There are many other examples. Two side notes on these articles. One, this kind of benchmarking has some shortcomings; Google’s increasing use of personalized search results means I might see CSO articles higher than you do. I’d be interested if somebody wants to note in the comments below their own results on these three searches. Two, you see that we frequently left some remnant of the original print headline intact, in case a print reader came to the site to find a specific article using a magazine headline search. Today I wouldn’t necessarily bother with that consideration.
At any rate, beyond these basic thoughts, a real answer to this “was it worth it” question requires some examination of the lifetime value of SEO. Rob O’Regan already wrote up the point of view of an ecommerce company. That’s worth reading. Here are a few observations based on my own experience and fixation with Omniture.
One, the more accurately your headline describes a niche content interest, the more likely it will be to meet a searcher’s information needs. That means incoming search traffic will be less likely to bounce and more likely to sign up for a newsletter and to have good time-on-page, time-on-site, and similar engagement metrics – discussed in detail in my previous column about bounce rate.
Two, an article that brings you just five new visitors per day means some 1,800 new people see your site each year.
Three, the ‘lifetime value’ of these efforts hasn’t been fully realized yet. If I get hit by a bus tomorrow, those articles will continue to draw and serve readers, for years in some cases.
What Would I Do Differently?
Given the same opportunity today, I would:
- Provide better training and oversight to the whole team on SEO principles.
- Make extensive use of keyword research tools to build a better SEO matrix and target my headline efforts more carefully.
- Insert more photos (see point three in the linked article) with appropriate labeling.
- Document the hours spent on these projects so I could answer “was it worth it” with some precision – or at least fake it better.