One of the most complex tasks facing an e-commerce professional is the development a complete attribution model capable of measuring and optimizing conversions. Many of us have invested considerable amount of time and money in the pursuit of finding the most effective way of measuring the contributions of the partners and suppliers to a given conversion in a manner that is which both fair and efficient. It seems there is always data lost somewhere along the path when measuring the contributions, this makes it difficult to be objective when trying to optimize investment or when justifying the expenses of a marketing campaign.<\/p>\n
In this article, we are going to profile the main attribution models as each one has its own characteristics however, before doing so we need to start at the beginning.<\/p>\n
What do we mean by attribution and what does it do?<\/b>
\nAttribution is a set of rules or factors that are taken into account when measuring the effectiveness of various different marketing campaigns, in the process we assign values to each channel based on their contribution to the end sale or objective, which then in turn helps us to optimize our online campaigns.
\nAttribution should not be confused with a payment model but should be seen as a way of understanding our users, the influence of different channels on them and therefore measuring their behavior. Finally it is a way of analyzing and maximizing the impact of our campaigns.<\/p>\n
The pros and cons of the most common attribution models:<\/b>
\nThere is a process of continuous development of a model which meets the needs of all the players involved, indeed, there are some basic models as well as some quite sophisticated ones available but they still have plenty of limitations when it comes to their accuracy and each approach has its advantages and disadvantages.<\/p>\n
Last click or last interaction:<\/b>
\nThis model attributes 100% of the value of the conversion to the last click.
\nPros: It is the most common model and thus the most widely used. Although it has its limitations we do at least have the certainty that something about the campaign or visit to the site, resulted in a conversion.
\nCons: it does not take into account the customer\u2019s journey or the previous navigations before reaching that all important last click.<\/p>\n
First click or first interaction:<\/b>
\nThis model does the opposite of the last click model by attributing everything to the first click or first interaction.
\nPros: none except that the records are useful to know which channels the user employed.
\nCons: It could well be a big mistake to attribute everything to an interaction that might have influenced the user\u2019s decision to buy and ignore all the other interactions (of which there could be many).<\/p>\n
Lineal:<\/b>
\nThis model gives equal value to all the channels.
\nPros: It is easy to apply, awarding everyone the same without any distinctions.
\nCons: By giving every channel the same value it distorts reality as it doesn\u2019t reward the real contribution of each interaction.<\/p>\n
Time decay: <\/b>
\nThis model takes into account all of the channels that have taken part but gives more credit to the closet in time to the sale or conversion.
\nPros: It gives more importance of the last interactions as they are nearest to the end sale and understands that it is those that drive or motivate the conversion or the sale the most.
\nCons: This approach is by no means perfect as the first or the first channels are just as important as the last one it is via them the user was introduced to the product or service.<\/p>\n
Position Base:<\/b>
\nThis model gives importance to both the first and the last channels giving approx. 40% to each of them; the remaining 20% is divided amongst the other channels.
\nPros: it appears to be fairest model given it rewards the first and last channels but does not forget the contribution of the other channels.
\nCons: It is remains an incomplete model as not every customer\u00b4s journey follows this pattern.<\/p>\n
Customized \/ Personalized attribution:<\/b>
\nWith this model it is the advertisers that assign the values to each channel or interaction by basing them on factors that are most valued to the advertiser and assigning each interaction a different value.
\nFor example we could assess the level of user engagement generated by each interaction using measurements such as length of visit, number of pages visited or the intention to buy (micro conversions).
\nPros: it is the most exact model giving a real and measurable value to each interaction.
\nCons: It is the most complicated of the models to use and it requires previous experience using the other models.<\/p>\n
The perfect model doesn\u2019t yet exist but perhaps a customized or personalized model is the most valid approach. A model where it is the advertiser who has to apply the rules and assign attributes to the different channels so when it comes to analyzing the customer\u00b4s journey it will vary for per brand or on a product basis. Each company using the model as a base would need to create their own version based on their own values but there remains two big handicaps when comes to measuring the results:<\/strong> the analysis of the model when the user is offline and that it doesn\u2019t take into account the use of multiple devices by the same user.<\/p>\n Moving forward it is essential to unite the online and offline interactions and to reconcile the use of different devices, as more and more customers and their interactions are lost in the gaps between them. The key to success is nothing that hasn\u2019t been done before but it remains a challenge: analyze, optimize, test, measure and then reanalyze.<\/p>\n","protected":false},"excerpt":{"rendered":" Multi attribution, the big challenge for e-commerce. One of the most complex tasks facing an e-commerce professional is the development a complete attribution model capable of measuring and optimizing conversions. Many of us have invested considerable amount of time and money in the pursuit of finding the most effective way of measuring the contributions of…<\/p>\n","protected":false},"author":6,"featured_media":5635,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"content-type":"","footnotes":""},"categories":[30,664],"tags":[478,479,480,693,723,724,725,726],"class_list":["post-5634","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-advertisers","category-tips","tag-analytics","tag-attribution","tag-multi-attribution","tag-ecommerce-en","tag-analitica-web-en","tag-marketing-en","tag-modelos-de-atribucion-en","tag-multi-atribucion-en"],"yoast_head":"\n
\nThe way consumers shop has changed, a customer can now see a product on an App go to the shop to check it out then return home to buy it online on the PC.
\nThe consumer is looking for the Omni channel without knowing, this means we have work fast to make technological innovations, the Omni channel buyer can provide up to 20% more conversions<\/strong> than the traditional consumer.<\/p>\n