Watching SEM Dials and Gauges

By Kevin Lee , December 19, 2003

For true performance, racecar drivers, fighter pilots, sailors, and search engine marketing (SEM) managers must navigate by their instruments. Like flying or sailing decisions, SEM decisions are balanced with experience and intuition. Even with autopilot and GPS, a human element results in smoother, more efficient execution.

The human factor allows strategy to be adjusted in subtle, yet critical, ways. Watching the dials and gauges takes into account subjective, external campaign factors. The best SEM managers know which dials and gauges to watch and how to adjust campaign objectives to meet overall marketing goals.

I discussed matching online-only data to an SEM campaign last week. My fellow columnist Frederick Marckini covered connecting offline sales to SEM using a case study. Let's continue looking at how both visible and less-measurable variables can be used to tune an SEM campaign for maximum profit.

No two marketers have the same business or objectives. Yet we can all learn from peers who are molding campaigns to their specific objectives. Let's delve into the on- and offline metrics you might want to take into account when piloting your SEM campaign.

• Offline conversions. Some offline conversions can be tracked using technology or couponing. However, tracking everything is impossible, so sometimes offline conversion data is extrapolated or anecdotal.
  
  
• Keyword volatility. Some keywords are volatile and require frequent automated checking to determine if the price/position parameters are within the optimal range.
  
  
• Traffic volume shifts. During some dayparts, your keyword mix will generate disproportionate levels of traffic (given a particular position). Knowing your traffic profile can help you make adjustments.
  
  
• Conversion rate by time. As explained in my dayparting column, a campaign may have a unique set of conversion behaviors in respect to the time of day a click is initiated. By understanding your specific profile, you can develop a strategy that takes advantage of changes in conversion by time.
  
  
• Conversion rate by day of week. Like dayparts, days of the week can impact conversion behavior. This holds particularly true in business-to-business (B2B), but also for many other service and product companies.
  
  
• Keyword mix. Once, only paid-inclusion listings resulted in traffic from a mix of different keyword phrases. Now, Overture and Google have broad and phrase match options. Your keyword mix for a single broad match listing can be quite diverse. Watch the mix for any keywords that negatively influence return on investment (ROI) and adjust your campaign appropriately. (Here's a how-to.)
  
  
• Competitive separation. Competitive separation (your ad is separated from a competitor by pages or ad units) is common in traditional media. In a paid-placement auction, you can only control position (and often not to an exact level). In search, competitive separation can impact conversion rate. Similar offers, such as companies with similar pricing or positioning, may run ads near one other. Although this may not dramatically impact conversion, it's a factor to consider when setting cost-per-order/-action (CPO/CPA) goals.
  
  
• Changes in effective CPO/CPA. Time of day, day of week, position, creative separation, and many other factors influence effective CPO or CPA. With an automated feedback loop, a campaign can be adjusted to reflect changes in effective CPO.
  
  
• Buying-cycle seasonality. Do searchers exhibit behaviors indicating they're earlier or later in the buying cycle at different times of the year? If a surge of searchers clicks on the technical specifications for a product, perhaps they're early in the buying cycle. That doesn't make them less valuable.
  
  
• Long-term buyer behavior. Are search engine-generated customers of equal value as "average customers"? Collect data to learn more about your customers. Use it to tune campaigns.

Some marketers looking for excuses to use subjective campaign management settings or budgets use terms such as "branding" and "awareness" to justify their decisions. Yet branding variables can be measured. SEM brand-lift studies may not be in every marketer's budget. If you can't afford a branding study, consider the keywords in your campaign, and if those keywords reach likely new customers or those already familiar with your brand.

SEM is a profession. Develop the experience to know when to change campaign settings based on data analysis. Combine this with adjustment factors that reflect your business reality. Sometimes you must make assumptions and back them up with less-than-perfect data. The bottom line will validate or disprove your hypothesis. That's why SEM campaigns are ever-changing.

Back to Article