In line with the growth of the French e-commerce (+12% in 2013) the pressure to improve margins keeps growing. In the past couple of years, French advertisers have redoubled their efforts to make acquisition investments as profitable as possible.

The key question is how to gain more by paying less, which translates into building out the most lucrative partnerships to the detriment of the less beneficial ones. It is a challenging situation and there are no easy solutions, but we French are well-known for our capacity to theorise and debate complex issues!

Many advertisers have invested in tools to help gain insights to aid attribution such as external analytical solutions or tag management. Armed with this analysis, the French merchants have typically developed four strategies:

  • Applying varying CPA and/or CPC commission rates. These are individually adapted based on the added value that each partner is perceived to bring to the business.
  • Deduplication. Allowing merchants to create a number of predetermined criteria to choose for each sale, which channel or partner should be rewarded.
  • Testing split commissions. Some merchants have set up a number of tests to see if it could be possible to split commission fees between several partners involved in the sale.
  • Maintaining state of play. Defaulting to a last-click CPA resolve being the least imperfect scenario.

Developing the different scenarios is one thing, but putting them into practice is often a lot more complicated. While some merchants are without a doubt very efficient in optimising the performance of the programmes they are running, others have tried hard to make improvements without seeing any real return on investment. In this context it is worthwhile to note the different impediments which have been identified in relation to each of the above:

1. The trouble with tag management

Although the tag management tools are quite sophisticated and offer a wide range of functionality, they can often take a long time to implement and even longer to work out how to make good use of. Additionally, as an industry we have a lack of neutral and constructive guidance to help retailers use these tools efficiently.

Finally, after rolling out such a strategy, some merchants have found that publishers have reduced promotion of their brand (or stopped the programmes entirely) due to a lack of transparency on metrics. Major French publishers will have their own ROI expectations to determine who they choose to promote. A lack of visibility and predictability in this area can challenge publisher KPIs and the easiest solution is simply to promote a different retailer.

2. Deduplication can give revenue a knock

Some merchants have implemented deduplication in a way which has resulted in CPA levers becoming an adjustment variable to CPC/CPM levers. In these cases, CPA publishers have reduced the promotion of these merchants, and as a consequence the revenue generated for these merchants had decreased.

3. Cashback and loyalty

Cashback and loyalty have proven impossible to attribute/deduplicate against. Examples of angry community members spreading complaints across the Internet, have undermined the intent of promotional activity leading to negative comments about the merchants in question.

As a result, certain cashback affiliates have ceased working on these programmes. Where cashback is a large part of an affiliate programme, its exemption from attribution roll outs has undermined the wider policy and even when it’s not a major contributor, there can be general ill feeling within the wider community about inconsistent messaging.

An interesting comparison made during these reflections was that the question of the right attribution model was a bit like imagining how to pay a football team incentivised solely on its performance.

Would it be correct to pay only the striker and ignore the rest of the team? Or on the contrary, would it be better to pay a goalkeeper or a defender depending on the goals scored by a striker (who, incidentally, may actually no longer get paid for goals at all)? 

We could push this to the extreme by imagining that perhaps part of the team is paid for the number of passes they play (a bit like a CPC model), and once their objective is attained, the players which used to be paid for their goals get nothing at all anymore. You think this sounds incredible? This is what happened in France.

After two years of debates and trials in France, it must be noted that the perfect model does not exist. Maybe in the end the winning model is a model where the retailer understands the intrinsic value of each partner and pays accordingly, attributing specific criteria (such as awareness, traffic, news users, transformation) to each one of them. A bit like a football team where each player is important but the coach understands exactly what it takes to get to the top of the league.