When SEM Needs to Back Off

As search engine marketers, we sometimes find it hard to slip the surly bonds of TCP/IP and realize that, as vendors, our primary goal is the overall success of the client, not simply the online success of the client.

That means making hard choices that may not always benefit us in the short term. In the long term, however, they usually do.

At times, it’s smart to objectively look at your recommendations and be prepared to step back if necessary.

The Pecking Order of Traffic Sources

When you argue for a certain type of change to site architecture, it’s smart to know at the beginning of the conversation how much clout you have. This is almost certainly determined by the amount of revenue the SEM (define) program generates for the company overall. Assuming all traffic converts at the same rate (I know, it doesn’t always), it boils down to a simple equation: as an SEM professional, your influence on site decisions is directly proportional to the percentage of traffic coming from search.

And that’s really the way it should be. Many big brand sites have invested significant dollars into SEO (define) and SEM, but less than half of their traffic comes from search. This is due to bookmarking, direct type-ins, aggressive nonsearch campaigns, predictive browser URL fields, and other behaviors that sidestep all your efforts.

Another argument that might stop you in your tracks is your client asking you to predict the ROI (define) of a suggested change. And they’ll even tell you what the “I” is. Most of the time, such a prediction is either extraordinarily difficult or downright impossible to make. And the last thing you want to do is oversell your recommendation so much that it becomes impossible to meet your predicted results.

If you’re absolutely certain your recommendations are valid, however, there are several options:

  • If a small test is possible (and requires less overhead to implement), try to gain buy-in for a trial run and extrapolate a predicted site-wide result.
  • Test the technique on another site and offer to report the results and predict a result from those.
  • Use your instincts to contrast your recommendation with past recommendations. For example, you might suggest that it will be more effective than the meta description initiative, but probably not as overwhelmingly successful as consolidating all the duplicate content you did the prior year. You may not be able to quantify your prediction in terms of visits or dollars, but such a comparison helps your client understand your prediction. And it doesn’t hurt that in doing so, you’re reminding them of two previous successes.
  • If all else fails, highlight the problem with a screenshot of an embarrassing SERP (define) and explain how your recommendation will alleviate the problem. Visual displays like this often help sway stakeholders, especially if you can highlight a competitor that doesn’t have a similar problem.

Best Practices, As Defined by You

If you come up against a business decision that won’t allow you to implement your version of best-practice SEO or SEM, don’t immediately consider them fools and yourself persecuted. Think strategically.

For example, I watch the automotive space very closely. For years, the Ford Mustang page has performed exceptionally well despite little effort put toward on-page factors.

Did I say “little effort”? That’s an understatement if you look at the cached page, because once you factor out the Flash, there’s nothing left.

Would that be your first choice for optimizing a Flash page? Probably not. Best practices? Of course.

There’s nothing untoward happening there. Ford has simply compensated for the on-page factors by working heavily on-site navigation and external linking. (Note, too, that the Flash site has a parallel HTML version that almost never shows up in SERPs.)

The moral here is that almost no SEO consultant would have vetted a page completely devoid of HTML content, but there it is. I like to imagine that during the meeting, Ford’s SEO person argued strenuously, but reluctantly gave in with a final warning: “We can do it, but you’ll pay five times for the extra link development what an HTML skeleton would have cost.”

And don’t write off the Mustang’s success to the “big-brand effect” or whatever other meme is floating around today. You might be surprised at the number of big-brand queries whose expected penthouse tenant is noticeably absent, with a scrappy upstart living in its place.

SEM companies need to admit that certain strategies will probably work, even if they’re not strategies they prefer to recommend. They just need to fully explain the conditions as well.


It’s alluring and a bit addicting to be the go-to person for SEM, but this frequently results in the need to prove your value through constant motion. In the long term, however, you’ve done a greater service to your client when you’re completely objective about the techniques you are, or aren’t, recommending. They’re paying you to make wise recommendations, and sometimes the wisest recommendation is to sit tight and store up resources until they’re really needed.

Join us for a one-day Online Marketing Summit in a city near you from May 5 to July 1, 2009. Choose from one of 11 one-day events designed to help interactive marketers do their jobs more effectively. All sessions are new this year and cover such topics as social media, e-mail marketing, search, and integrated marketing. Register 30 days in advance and get a $40 discount!

Related reading

Brand Top Level Domains