Every marketer at some point has faced these difficult questions:
- Should I bid on my competitor’s brand terms?, and
- How should I react when they start bidding on my own brand terms?
We get asked these questions all the time. The strategies around competitor bidding are diverse and vary depending on whether the goal is performance based or some other strategic or branding goal. In this post we will focus purely on performance considerations which means we are asking: what bidding strategy will maximize my profit? We will illustrate the main arguments through an example and show the results of a test we did with one of our clients.
Competitor bidding theory
In the following example Advertiser A (AdvA) and Advertiser B (AdvB) are bidding on their own brand terms and each others. For the sake of simplicity we will assume there are no other participants in the auction. So AdvA will bid on “AdvB” keywords and AdvB will bid on “AdvA” keywords. This creates a scenario where AdvA acquires some traffic from AdvB and at the same time loses some of its own brand traffic to AdvB. In addition AdvB, by bidding on AdvA keywords, is increasing the CPCs of AdvA for the totality of its traffic. Furthermore we assume that both advertisers retain the top ad position when bidding on their own brand terms.
From the perspective of AdvA, it is “stealing” some traffic from AdvB brand keywords, while having its own brand clicks become more expensive. To understand this more in depth, we will look at the math using the following denotations:
Traffic_{B}^{A} denotes traffic from AdvA brand keywords acquired by AdvB. In that case, the profit obtained by AdvA from its bidding on AdvB brand keywords is given by:
Where, RPC_{A}^{B} is the revenue per click of AdvA on AdvB brand keywords traffic and CPC_{A}^{B} is the cost per click of AdvA on AdvB brand keywords traffic.
The additional cost of AdvA due to AdvB bidding on AdvA brand keywords is given by:
Where \Delta CPC_{B} is the delta cost per click of AdvA due to AdvB bidding on AdvA keywords. Similar equations can be written for AdvB. With those two equations, we can start making some assumptions on the different scenarios.
If AdvA and AdvB are two big advertisers with well-established brands, then Traffic_{A}^{A} \gg Traffic_{A}^{B}. In that case, it is most likely that \Delta Cost_{A} > Profit_{A}^{B}, which means that the extra cost to AdvA due to AdvB bidding on its keywords will be larger than the profit made by AdvA bidding on AdvB keywords. Using some typical values of, say, a conversion rate of 3% on competitor keywords, a revenue per conversion of $50 and Traffic_{A}^{A} \approx 10 \cdot Traffic_{A}^{B}, the ratio \Delta Cost_{A} / Profit_{A}^{B} \approx 4. This means AdvA is losing four times more profit from the competitor bidding on their keywords compared to what they are making by bidding on their competitor keywords. The interesting conclusion is that if AdvA and AdvB are of similar size and market share, it is better for both advertisers to stop bidding on their competitor brand keywords in order to maximize their profits.
This is a particular case of the prisoner’s dilemma as depicted in the following diagram. The diagram shows the payoff matrix, which in this case denotes the profit of AdvA and AdvB. In the case that both AdvA and AdvB bid on their competitor brand keywords, both will make a profit of 1 (normalized), if AdvA bids on AdvB keywords and AdvB does not bid on AdvA keywords, AdvA makes 5x profit, while AdvB makes zero. The optimal strategy is achieved when both stop bidding on their competitor keywords, as the maximum combined profit is achieved.
Competitor bidding in practice
Bidding between equal sized competitors
We wanted to try this strategy with one of our clients that met this scenario. So, after we finished the calculations depicted by the equations, we noticed that it would be five times more profitable to stop bidding on a competitor and “persuade” them to stop bidding on our client in order to achieve an optimum profit scenario. Figure 2 depicts the test and the results as a function of time. In this case we implemented the strategy for AdvA. The motivation started as AdvB started to bid more aggressively on AdvA keywords and hence increased the Brand CPC’s of AdvA as shown. We decided to stop bidding on that competitor (AdvB) and waited. Following a 2-week period of time, AdvB stopped bidding on AdvA and henceforth CPCs of AdvA went down by almost 35%, and increased the overall profit of AdvA.
Bidding between different sized competitors
The math changes when you are either bidding against a much bigger competitor or a much smaller one. Let us assume that your competitor is so small it gets no brand traffic at all. In such a case there is zero potential to gain clicks from the competitor’s brand terms and only the downside of increased CPC from them bidding on your brand. As before it is very much in your interest for neither company to bid on each other.
However, from the small competitor’s perspective, with nothing to lose, it does makes sense for them to bid on your brand. Does this happen in practice? I have spent a few moments searching for some of the top car rental companies. Sometimes the competitor ads are at the bottom of the search results or on the second page, but in every case ads from small competitors show whilst the larger ones are absent. Now you know why!
** Profit metrics are normalized for privacy purposes.
*** In the following weeks we will be rolling out a simulator to simulate competitor bidding scenarios of your own, stay tuned
