Judging PPC Performance: Focus on Conversions, Part 2

How do you know whether your keywords are performing well? Should you delete or pause under-performing keywords – or simply adjust their bid prices?

These are critical questions for all PPC advertisers. And it’s tough to answer them generally, since the answers may vary depending on the advertiser’s objectives, and the speed with which impressions, clicks and conversions accumulate.

Before we dig into the details, let me underscore the most important point about keyword and ad optimization. As I mentioned last week, decisions about keyword and ad performance should mainly depend on conversion performance. (If you’re not tracking conversions, stop reading and go do it now — here’s a good place to start.

Many advertisers make the mistake of making keyword and ad decisions based on impressions, clicks and click-through-rates. Trust me on this one: frequently, an ad or keyword seems to perform well according to these metrics, but fails to get as many conversions as other keywords or ads that get.

Here’s some good news for advertisers concerned about time management: if you read my previous column about creating tight ad groups, you won’t need to worry about individual keyword bids. The small number of keywords per ad group will garner very similar performance characteristics. So you can concentrate on determining the ad-group-level bid price.

Another advantage to creating small, tight ad groups: you won’t need to worry about poorly-performing keywords dragging down quality score for the ad group. Small, tight ad groups get the best possible click-through-rates, and hence good quality score is practically guaranteed. In such ad groups, there are no poorly-performing keywords – just keywords that get an acceptable number of clicks and conversions, and others that accumulate clicks and conversions less quickly. You won’t need to adjust individual keyword bids, nor worry about pausing or deleting underperforming keywords.

But if your ad groups are imperfect in this respect, how should you make keyword decisions? Our simple rule of thumb is: you shouldn’t make a decision about whether to adjust bid prices, or delete a keyword, until the keyword has accumulated at least 30-100 clicks and 30-100 conversions.

The exact number of clicks and conversions necessary to make decisions depends on the velocity of your ad groups. If your ad group accumulates clicks and conversions quickly (say, hundreds or thousands of clicks and conversions per day), you should use the high end of that range. If the ad group accumulates clicks and conversions more slowly — hundreds of clicks and conversions per month, for example — then you should use the low end of that range.

Purists might say that 30 clicks or conversions are too little data to make decisions based on statistical validity. While that’s indisputable, for many advertisers it’s simply impractical to wait the weeks or even months before enough data accumulates. Certainly there’s a risk that an incorrect decision will be made — e.g. a keyword may be shut off that might perform well if left to run for a longer time. But for many advertisers, that risk is outweighed by the risk of failing to make any decision.

I expect this viewpoint to be controversial. Let me know what you think.

In a nutshell, here are the keyword guidelines for busy advertisers:

  1. If a keyword has accumulated more than 500 impressions, but no clicks, pause or delete it. It is adversely affecting the ad group’s quality score.
  2. If a keyword has accumulated 150-200 clicks but no conversions, pause or delete it. It’s hurting your ROI.
  3. If a keyword accumulates even just one conversion, it’s golden. Watch as it accumulates 30 or more conversions — if cost-per-conversion is above your target, lower the bid price. Likewise, if the cost per conversion is lower than your target, you can afford to increase the bid price.

Next Monday I’ll resume the discussion of PPC bidding strategies; how and when you should adjust keyword and ad group bid prices. As usual, let me know your comments and questions via the feedback form below.

Listen to David sharing his PPC tips in a live Search Engine Watch Webcast, “Profitable PPC: The Fundamental Secrets,” on October 22, 2008.