Thursday, March 26, 2009
The Quality Volume Divide
It is a problem old as time in the performance marketing sector, one that shows itself much more clearly and painfully in the world of online lead generation. The challenge to produce high quality with high volume. Ask almost any advertiser who has at least a modicum of experience in online advertising, and growing volume while maintaining quality will rank high among their challenges and frustrations. The unfortunate truth is that after a certain point, quality starts to degrade. What exactly is quality, though? The reason that we use lead generation as an example is that quality is easy to explain. Let's use an auto insurance offer running on a CPA network as an example. The offer looks to get users to enter their information to see if they could lower their auto insurance payments. When a user enters their information, the network receives credit and they then credit the appropriate publisher. The person buying the lead receives no money from the user filling out the form only from the percentage of users who then go on to purchase a policy. The higher that conversion rate, the higher the quality. If more people convert from lead to policy, the lead buyer can afford to pay more. If fewer people do, then they will have to lower the payout in order to continue covering the cost of buying the leads.
In the optimal scenario, conversion rates start out profitably and even increase over time as both sides optimize. At a point, though, especially in the optimal scenario where the advertiser sees good returns with good volume, they will want more. Two things start to happen here. The first is often counterintuitive for the advertiser, and we called it the price fallacy in lead generation, namely that more volume comes at a higher price. What the price fallacy fails to capture is what, more often than not, happens to quality. Once under the gun to get more volume, when suppliers attempt to do so, they end up succeeding but at the expense of the initial quality.
As the illustration above shows, the optimal phase sees volume growing with quality remaining above the break even. When the two parties switch into the forced growth phase, volume continues to increase (often it increases at a slope higher than the initial growth), but quality starts to slip. More quickly than expected it goes from the advertiser having a positive yield to a negative one.
Quite a few explanations exist for the quality-volume divide. One of the more straight forward revolves around intent. Only so many people have a given interest in a product. B2B marketers deal with this issue all the time. For some high dollar, super complex sales, e.g., a multi-million dollar database configuration, there just aren't that many people who could be buyers. With auto insurance, the number is fortunately much higher, but it's not infinite. Different traffic channels have different levels of intent - search is not surprisingly higher than co-registration. For many verticals, there are only so many keywords available. To get in front of more users it means trying other avenues - email, display, contextual ads, etc. Each one of those will have its equivalent of head users and tail users - sites / placements where users who click will have an interest as opposed to someone who places an ad on Facebook saying, "Find out how much it is to insure a Ferrari." Each incremental step in obtaining more traffic tends to come at the expense of the intent of the person who views / clicks / converts on the ad. Here is what it looks like plotted.
Saturation also plays a role. At some point, an advertiser will simply have reached the vast majority of potential users for the product. The problem is that they want to grow. They don't want to settle for the same amount, and it pushes them to continue trying even at the detriment of their quality. It's not that higher volume and good quality can't go together. It's all about the incremental lead in the growth phase. In the forced growth phase, instead of the next lead converting at or near the previous lead, it keeps slipping. Good leads still exist, but they get lost among the lower quality ones. It's a problem which, if solved, will mean incredible gains but it currently falls outside the skill set of both the lead buyer and lead seller. Each can get better - they can implement various levels of verification (quality, scoring, call centers) but to get really good, means each getting away from what they do best. Buyers and sellers get closer, but they will never close the divide fully. It's too complex, too distracting, and not urgent enough for them. All of which means one of two things will happen. The divide will come and bite everyone in the behind and/or someone else will come along, solve it, and do very well. A note of warning though, there is another reason why no one has done so to date. It's anything but easy.