Managing Tail Terms: Balancing Risk & Reward

The savvy search marketer always explores tail terms to improve overall performance. If done right, tail terms can convert better than the head and can also be cheaper in terms of Cost Per Click (CPC). Hence, effective tail management lifts the performance of the overall campaign. When exploring the tail, marketers are always faced with […]

Chat with SearchBot

The savvy search marketer always explores tail terms to improve overall performance. If done right, tail terms can convert better than the head and can also be cheaper in terms of Cost Per Click (CPC). Hence, effective tail management lifts the performance of the overall campaign.

When exploring the tail, marketers are always faced with the dilemma of risk and reward. Bidding on underexposed keywords aggressively might bring in a lot of conversions and uncover gems in your campaign. However, you could also end up spending a lot without any conversions. The question then is how to identify potential candidates while mitigating risk.

The most effective way of managing the tail is to use an algorithmic approach where you can use a finite mixture and confidence models. This is a technology-based approach that scales very well and has done very well for the clients I manage. However, if you are a small advertiser without access to sophisticated technology you are probably looking for a simpler heuristic method.

zero-imp

Consider a campaign of 10k+ keywords as shown above. Among these, you have enough data to determine if the keywords are good or bad for 1,470 of them. These are high confidence keywords where we know exactly what to expect. However, the majority of these, 8,908 to be exact, are lower confidence keywords where we do not have enough data. Of those, 3,000 plus keywords have seen page one and, hence, have been given a chance. However, 5,797 keywords have not seen page one and, therefore, are potential gems. Marketers should focus on these keywords.

However, here comes the risk part. If many of these keywords are already bid over the average campaign CPC (here $1.07), then bidding them higher would increase your risk. You should focus on the low bid keywords as they have not been exposed and have a lower risk potential. These keywords (marked in red) are your best learning opportunity. In these keywords, several of them have not even seen an impression. So, in a sense, these keywords should preferentially be bid up and monitored. Once the low bid, low position keywords have been exposed, you can move to the next tail opportunities such as low bid-higher position and high bid-low position keywords.

If you are interested in adopting such an exposure strategy, I recommend the following steps:

  1. Calculate the average clicks per conversion on your head terms. Preferably, you can calculate it for all your keywords and have a distribution. Determine the threshold below which you don’t have confidence on a keyword’s value. Lets say you find the mean clicks per conversion is 100 with a standard deviation is 30 clicks. You can assume that 100-2*standard deviation as your threshold i.e. 40 clicks or less form the low confidence set.
  2. Next, determine the average CPC of the campaign. This is your bidding threshold.
  3. Find those keywords that have a lower than average CPC with less than 40 clicks and those that have not seen page 1 in recent history. This set of keywords is the low risk, higher potential set.
  4. Incrementally bid these keywords and monitor their performance.
  5. You can repeat this exercise every week to two weeks but use a longer window when evaluating tail term performance. My research has shown that only half the keywords that get clicks in one month get clicks in the consecutive month. So use a long evaluation window.

This strategy is a useful starting point when attempting to maximize tail term performance. Once you build scale, you can start thinking about automated approaches.


Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.


About the author

Siddharth Shah
Contributor
Siddharth Shah is head of web analytics, digital strategy and insights at Adobe. He leads a global team that manages the performance of over $2 BN dollars of ad spend on search, social and display media at Adobe.

Get the must-read newsletter for search marketers.