Multi-touch attribution is a modern method of determining the value of each customer touch point leading to a conversion using Big Data approach. It includes the full digital path-to-impression and click stream cross-device data, TV and Radio advertising data and incorporates Creatives and Special Offers data.
Multi-touch attribution models gives the marketer unbiased view on performance of marketing vehicles quicker to standard Marketing Mix Modeling and Digital Attribution studies.
Digital / Multi-touch Attribution
Digital / Multi-touch Attribution is a method of distributing conversion credit among marketing channels (marketing touchpoints) along a customer’s journey from star do final conversion. Marketing practice splits Digital Attribution approaches into Rule-based/Heuristic attribution and Algorithmic attribution.
Rule-based attribution approach is humanly defined in arbitrary way. The most popular models are
- Last Click (or Last Non-direct Click) – the last click (or the last non-direct click) has attributed 100% of credit for the conversion. This method simply ignores all previous user’s contacts with brand communication, that might have also contributed to the conversion.
- First Click – the first click has attributed 100% of credit for the conversion which ignores all other channels and rewards only opening marketing instruments.
- Linear/Even approach – each touchpoint alone customer’s journey has identical portion of credit and channels with highest frequency are rewarded the much. Very simplistic and arbitrary model.
- Position based, Time decay and other rule-based – all of them create an arbitrary split of credit for the final conversion. Some of them in specific cases might be useful in analytics process. However none of them can be freely use to produce fair distribution of credit by channels’ performances.
Algorithmic attribution is about leverage on advanced statistics and machine learning algorithms determine (to without any additional assumption) the impact of marketing channels along a customer’s journey toward conversion. It leads to a better understanding of marketing-campaign effectiveness. Algorithmic attribution requires a good, standardized data source where complete information about campaigns is stored. The most popular approaches are Markov models, Shapley’s regression method and Survival models.
Web analytics providers like Google or Adobe can deliver specific reports containing all click-based customer journeys records. For marketers and brands focused on performance marketing and click-stream conversion processes this may be a great data input for algorithmic attribution model. However web analytics tools use only converted paths to help determine the results, meaning company might only be using few % of the data collected as in majority of cases conversion rate is in the range of 1-10%.
Algorithmic attribution on Click-stream data
More and more companies use algorithmic attribution models based on Google Analytics data to understand the real value (effectiveness) of every marketing channels on customer’s journey towards final conversion. Advanced algorithmic attribution models measure the impact of each digital channel that influenced conversions. With attributed media spend we can create the ROI ranking per digital investment where to invest your digital marketing budget to optimise campaign performance.