The Merkle Q4 Digital Marketing Report (DMR) was released earlier this quarter to highlight trends our client base observed from October through December of 2016. In the Comparison Shopping Engine (CSE) section there was one major theme that deserves a deeper look: the decline of Connexity from both a spend share and revenue growth perspective.
The Cold, Hard Facts
As a refresher, the DMR called out the below performance trends for Connexity:
- Connexity has struggled to maintain spend share since accounting for 50% in Q1 2016. They are currently holding steady at 29%.
- Revenue from Connexity ads declined by 30% YOY, while revenue driven by ads on eBay Commerce Network rose 30% YOY.
- eBay Commerce Network significantly outperforms Connexity on conversion rate across most product categories.
Our Q4 data very clearly illustrates that, between the two major Comparison Shopping Engines, ECN is substantially outperforming Connexity.
What Is Driving Connexity’s Decline?
The overall reason for less investment in Connexity is that retailers aren’t seeing efficient returns from the engine. We are usually able to take a client’s goals regarding ad spend to cost ratios and adjust bids or product mix on a given platform to maximize revenue.
If a product, keyword, or product group is performing inefficiently, we look to reduce bids to a level that works with our expected sales-per-click (SPC). Reduced bids should lead to lower cost-per-click (CPCs), lower click volume and, as a result, lower overall ad spend.
Connexity’s platform bucks this trend in that lower CPCs do not necessarily mean lower costs or better efficiency. The data below illustrates the correlation between weekly CPCs and efficiency for 19 Merkle advertisers during Q4 2016. A value of 1 indicates perfect positive correlation, meaning CPCs and ad-spend-to-cost (A/S) increase together in a linear fashion. A value of -1 indicates negative correlation, or that A/S decreases (efficiency improves) linearly when CPCs rise.
Overall we’d expect to see a largely positive correlation, with more aggressive bids and CPCs leading to a higher A/S, meaning less efficient performance. Exceptions might be if a retailer is already showing very prominently, in which case higher bids won’t add much incremental traffic, or if their bids are so low that increases won’t get them anywhere close to the first few pages of results. In our client sample of 19 retailers, correlation figures were all over the place with eight positive, nine negative, and two right around 0.
For comparison’s sake, here is what the correlation data looked like for eBay Commerce Network when looking at the clients active on both platforms. All of the values are positive, with just over half at .5 or higher.
The results for Connexity are surprising given what we see across other platforms. One reason we believe this occurs on Connexity is that lower bids can result in ads no longer qualifying to appear for higher value searches on higher value domains in the Connexity network, which require higher bids. Thus, by decreasing the bid for a product, advertisers end up reducing the traffic that drives the most value, and are left with less valuable traffic requiring even lower bids.
This situation creates a downward spiral effect for many programs. Advertisers decrease bids to get more efficient, but that chops off the best performing traffic and forces greater pullbacks. Eventually one is left with just a handful of products at low bids in order to keep the program afloat.
In many cases, the only viable options are to 1) reduce bids so low that even a substantial increase in traffic would still lead to lower overall ad spend or 2) pause the program. As such, many of our clients are steadily reducing Connexity investment, and some are walking away from the engine entirely.
Ways to Right the Ship
There are several improvements Connexity could make to help alleviate some of the recent performance issues we’ve experienced. One of those options would be giving advertisers the option to stop serving traffic on underperforming domains, a feature similar to what Bing Ads offers, which gives analysts the opportunity to eliminate underperforming websites from their advertising mix.
If that’s not possible, a traffic estimation tool similar to what Google provides would be helpful to understand how a bid change will impact expected overall ad spend.
Another change that could improve performance would be updating the current bidding system. Connexity removed the Target Bid cap in June 2016, giving them the freedom to charge above your target bid (what you’d normally think of as your Max CPC on other search and comparison shopping engines) for “higher quality clicks”.
Prior to the June announcement, CPCs were capped at 30% above your submitted bid, and the idea behind this change was to give Connexity more flexibility to help advertisers chase higher-quality traffic in addition to discounting poorer performing traffic. While the concept is good, we would prefer that Connexity use the specific bids calculated using our proprietary internal systems.
The 30% bid increase might sound familiar. It reads similar to Google’s enhanced CPC program, which “adjusts your cost-per-click to help maximize conversions” by raising your max CPC bid up to 30%. One major difference is that Google offers this as an option, while all traffic not bid at Connexity’s Single Rate pricing is subject to increases at its discretion.
The Single Rate bidding model, while different from the Targeted Bid, also has an element of `black box algorithm’ to it. Under the Single Rate model, advertisers submit a bid of at least $4 in order to make a product eligible to serve and the “Smart Pricing system chooses the appropriate CPC based on the quality of the lead.”
While observed CPCs have never climbed to $4, the lack of control is troublesome and, depending on demand for a given product at certain times of the year, could lead to CPCs beyond what you’d otherwise be willing to pay. Additionally, any bid below 5 cents will be automatically increased to the Single Rate $4 bid.
Recommendations for the Immediate Term
There are a few boxes advertisers can check to best set themselves up for success on Connexity in its current state:
- Implement either the conversion pixel or daily sales report exports so Connexity can see the return driven by certain ads/partner sites and discount traffic as warranted. More details can be found by looking in the Connexity UI or talking to your account manager.
- Monitor bids closely and react quickly to any spikes in click traffic that occur from exposure to different partner sites.
- Analyze which products, brands, and categories perform best for your site and consider including just those items in your feed.
- Set a default bid for each product subcategory in the Connexity UI to create a safeguard in case you submit a blank bid in the feed. If you submit a blank bid and there is a subcategory bid present, Connexity will use that subcategory bid. This can help prevent seeing substantial CPCs and click traffic when it was not intended.
It will be interesting to see if this trend continues throughout 2017. If Connexity does not change these aspects of their program, we expect to continue seeing year-over-year declines in both spend share and revenue, with ECN emerging as the only major player left in the CSE space.