Go to Top

Buyer Beware: More Than 50% of CSE Sales Attributable to Google Shopping

I want to address an issue that I don’t think gets enough, if any, exposure in this industry; Comparison Shopping Engines (CSEs), and the dishonest practices which some CSEs employ.

When I say CSE, I am not referring to Google Shopping, Bing Shopping or Yahoo Gemini Product Ads. We do not classify PLA networks such as Google Shopping, Bing Shopping or Yahoo Gemini Product Ads as CSEs, as there are distinct and powerful capabilities that PLAs possess which CSEs do not provide – such as absolute control over bids rather than using an arbitrary rate card, control over which search queries trigger an ad, they are generally more transparent, etc.

There is a dirty little secret which many, though not all, comparison shopping engines have. A secret which enables them to inflate their own performance reporting, making them appear more valuable on paper, allowing them to continue getting paid. This secret comes at the expense of the retailer – the retailer continues to feed a system which they believe generates incremental profit for their business, when in reality they are being taken advantage of.

I’ll explain with a story.

We began working with a new client, managing their PLAs, earlier this year. During our initial evaluation, we found that their Google PLAs were being tagged several different ways, with sales attributed not just to Google Shopping, but also to numerous CSEs. This essentially fragmented the measurement of their Google PLA channel. Upon further analysis, we found the CSE-related tags on their Google Shopping ads were not coming from their product feed, but rather from the comparison shopping engines which they were working with.

After some discussion with a CSE rep, they admitted that “almost all CSEs drive a majority of traffic through partners, including Google PLAs” though they “generally don’t break this number out.” Essentially, when this retailer signed up with these CSE networks, they were silently opted into Google PLA management through the CSE network – a duplication of effort as they already had dedicated PLA management. Not only does this practice provide zero visibility into the true performance of the CSE network, but it also forces the retailer to pay the CSE to manage PLA traffic under their arbitrary rate card, and impedes the retailers’ ability to manage their own PLA channel optimally. Meaning not only is the retailer paying the CSE a premium to access a network which they are already on (Google Shopping), but they also lose the ability to accurately hold the CSE accountable for their performance.

Think about that for a second. Would you deliberately give up margin to a CSE by managing Google PLAs through their network, on their rate card? Of course not.

While some CSEs refused to disclose exactly how much of their reported revenue came from their network and how much came from Google PLAs, one rep did state that for this particular retailer “over 50% of the sales were coming from Google PLA.”

Another CSE rep, from a different network, provided us with a bit more data, giving us a split of sales and expenses between their network and Google PLAs – which they managed but attributed to their network. Here’s the results for the period measured:

CSE Performance

This particular CSE reported $7,738.86 in total revenue on their network at a cost of $1,841.41 during this period, bringing A/S to 24% on average. With a 70% COGS/VOH expense, that leaves this retailer with $480.25 in profit, or a 6% margin. However, when performance was broken down further to reveal Google PLA performance vs the true performance of the CSE network, we found that the entirety of the profit driven for this retailer came from Google PLAs, while the CSE failed to produce any profit. Worse yet – what is the opportunity cost of allowing a CSE network to manage your Google Shopping channel?

If you’re currently working with any CSE networks, ask them if they have opted you into their Google PLA management service. If so, find out the true split of performance between their network and Google PLAs.

 

, ,