The Trouble in Forecasting Organic Search Results

When an organization negotiates to retain an SEM (define) firm, the big questions inevitably come up: “If we go with your company, what type of increase will we see? What will happen to sales? How much will traffic go up?”

For any business ready to plunk down a sizable portion of its marketing budget for organic search marketing, these are legitimate questions. In fact, the questions apply to nearly any purchase. How many calories does this have? What are the odds that I’ll walk again after surgery? How long will this new roof last?

In a metric-crazed industry like SEM, it’s odd that most SEM firms throw up their hands and say, “We’re not really sure. It depends.”

A Common Industry Dilemma

In his “Oilman” blog, Range Online’s in-house SEO guru Todd Friesen recently wrote an interesting post about the difficulties of forecasting organic SEO results. It made me feel better knowing it’s a pretty common problem across the industry.

I use (and recommend) several keyword prediction programs, which share one shortcoming: an inability to predict monthly query volume based on a very narrow share of total online searches. Keyword tools are very good at showing two terms’ relative popularity, but I think it’s nearly impossible to predict how many people will search for a given term by extrapolating 0.0065 percent of all searches out to universal proportions.

Consequently, to echo Friesen, even if we could guarantee a number one ranking for a set of terms and even if we knew the percentage of clicks that went to rankings one through five, we still couldn’t assume we’d know the grand total of people searching for a unique term.

Progress Is Not Necessarily Linear

“We made three out of your four recommendations. Shouldn’t we see 75 percent of your projected improvement?”

Many clients don’t understand, often because it’s not fully explained by the SEM firm, that various site recommendations and their resulting benefits aren’t always independent or linear. In other words, if four recommendations (A, B, C, and D) to a site architecture constitute 100 percent of our recommendations, making three of the four does not mean the company will reap 75 percent of the ideal site improvement. They might be completely reliant on one another. If change A means breaking down a crawling obstacle to deep pages, but changes B, C, and D are enhancements to deep pages, then a site will see no meaningful improvement unless change A is made.

The Paradox of Experience

Making matters worse, there’s often an unfortunate paradox in the relationship between clients’ SEO/SEM awareness and an SEM firm’s ability to help them. Businesses ignorant of SEO principles often have the most to gain from a good program because their sites possess such poor architecture. Yet knowing so little about SEM, these clients are difficult to find and it’s sometimes hard to explain the potential benefits.

Similarly, SEO-savvy clients have likely done their homework, implemented some sound changes on their own, and may have only a few issues that must be ironed out.

What’s the Answer?

About the best we can do is use past performance of a variety of sites to determine a range of potential improvement. In other words, find sites whose baselines best match the current client’s initial circumstances, and show the various results of no, partial, or complete implementation of the SEM firm’s recommendations. In determining that magical “initial circumstance,” take into account factors such as:

  • Percent of the site currently indexed versus total theoretical page count.

  • Total number of issues found during a comprehensive site analysis (with issues weighted according to severity and difficulty/time spent in implementing).
  • The site’s current list of referring keyword phrases compared to a list of potential short- and long-tail phrases for that industry.
  • Percent of existing content that requires minimal changes made. This could be as simple as determining the percent of pages with identical page titles and meta descriptions or as complex as a full content audit.
  • Amount of new content needed to rise to the level of minimal keyword demands for the site’s vertical market.
  • An honest assessment from the client about its ability to get changes implemented, as well as a rough timeline. This can be difficult to obtain, because most clients are overly optimistic about their ability to get changes pushed through various layers of bureaucracy.

In graphing these values, present the resulting grid as a visual representation of potential improvement, not guaranteed improvement. An SEM firm’s recommendation document, no matter how brilliant, will never result in search gains until it’s fully and correctly implemented.

If you’re working in-house, don’t despair. Whereas SEM firms might have many campaigns under their belts to use for data, you can too. You can optimize smaller portions of your site and divide them into topical areas, 10-page groups, or whatever is convenient. Call each segment a “campaign” and use the same criteria above to gauge your success over time.

Reap What You Sow

Early in the sales stage, a cursory glance at a site will almost never reveal the full potential of improvements. If a firm promises a certain improvement in results without digging into a site, chances are they’re aiming purposefully low to hedge its bets. This isn’t necessarily bad or dishonest, but it’s important to recognize when evaluating potential vendors.

Want more search information? ClickZ SEM Archives contain all our search columns, organized by topic.

Related reading