Knowing which channel triggered a conversion is crucial. It allows you to confirm the relevance of your advertising investments, and to optimise your media budget.

To achieve these objectives, you generally have two levels of analysis at your disposal:

1. Attribution: which aims to calculate the profitability of your channels; and
2. Contribution: which seeks to ascertain the role of your levers in your conversion paths.

Are these two approaches - attribution and contribution - fundamentally opposed, or do they complement each other? Which model best reflects the reality of your media mix?

‍

What are traditional attribution models?

‍

Imagine you're an acquisition manager with a significant advertising budget. Your campaign includes levers such as display, SEA advertising and affiliation sites.

With a large budget to justify to management, your priority is to be able to measure the performance of each of these channels (conversion rate and impressions), and value them accordingly.

To help you calculate your KPIs, you can use single-touch attribution models that attribute conversion to a single channel. These enable you to value your channels at the beginning of the journey (first-click) or at the end of the journey (last-click). Based on the conversions attributed using this rule, you can then calculate the conversion rate for each of your channels.

This single-touch attribution approach is well known in the market and is included in common analytical tools.

The advantage of this method is that it can be easily implemented and understood by your teams and management. It also enables you to simply and effectively remunerate your various partners and advertisers (affiliation agencies, comparison sites, advertising space, etc.). By attributing your conversions to a single channel, you can deduplicate them. In other words, you only pay for the partner to whom the conversion has been allocated.

However, as we can see quite easily, these traditional attribution models have certain limitations. Their main drawback is that they fail to do justice to all the channels that indirectly influence your conversion.

In this case, by using the last-click model, you are clearly enhancing the value of your channels that come into play at the end of the journey (such as branded SEA and affiliation), at the cost of other potentially equally influential channels in your mix (such as display acquisition).

This is why the contributory approach to these models is an essential alternative.

‍

The benefits of the multi-touch approach

Now, let's take the example of a football team that you would be coaching. To maximise your players' performances during the next match, you would like to assign the role that best corresponds to each of them on the pitch: instigator, passer or striker. You'd also like to make sure that they fulfil the role you've given them within the team.

It's exactly the same with the channels in your media mix. Each of your marketing channels fulfils a role in your conversion path, whether it's initiating, relaying or triggering conversion. To ensure the quality of these channels, you need to adopt a two-stage approach to your marketing mix:

1. For your channels that don't yet have defined functions, you want to identify their role within your conversion path, such as the Facebook or Snapchat campaigns that you've launched for the first time.
2. For channels whose role has already been defined, you want to make sure that they are fulfilling their function. For example, you want to be sure that your display acquisition channel is fulfilling its role, which is generally that of instigator, and your retargeting lever that of striker.

Based on this analysis, you can then create a contribution model (multitouch) that accurately reflects the performance of all your channels. These models enable you to give weight to each of your touches and optimise your acquisition budget accordingly.

Depending on your knowledge of your marketing mix and your objectives, you can choose between :

• Linear: if all your touches have an equivalent influence
• Increasing: if you have a high level of brand awareness, a high average shopping basket and you want to absorb the cost of your branding campaigns.
• Decreasing: if you have low awareness and want to capitalise on your initiating levers
• Parabolic: if you want to favour your initiating and scoring levers
• Algorithmic: if you simply want to use a complex model with recognised algorithms

The benefits of this contributive approach compared with monotouch models are obvious: you apply the analysis method that best matches your industry's typical conversion paths. That way, you gradually get closer to the reality of your customers' journeys.

‍

Why should you choose algorithmic attribution models?

If you want to go further in the in-depth analysis of the performance and role of each of your channels, you can adopt algorithmic attribution technologies.

With these models, you get a more objective picture of your advertising investment's profitability, and you can revalue the channels that contribute most.

‍

Our solution: Combine attribution and contribution models

Our Attribution solution is based on a pioneering and innovative methodology that analyses all your conversion paths, the interactions between your levers and their degree of autonomy.

It allows you to choose from a wide range of attribution and contribution models to analyse your mix (position, Markov, Shapley, etc.), supplemented by the our home made model for greater objectivity in your decisions.

In addition, the Attribution solution includes other features to help you manage your media budget:

• ROI measurement by combining sales and costs across all channels (including TV and Social...).
• An omnichannel view, to capture the impact on media, the website and in-store visits —including the Research Online, Purchase Offline (ROPO) effect.
• An ergonomic, customisable dashboard that lets you assign weights, roles and priorities to all the keys.

‍