Why it is essential to attribute on the basis of individual lifecycle

Practical guidance for modelling full customer lifecycle attribution

With ongoing pressure on marketing budgets and a burgeoning array of digital marketing channels and choices, attribution is a hotter topic than ever.

‘Last click’ is the most common approach; attributing 100% of a sale’s value to the last or converting click.

‘Last click’ is a flawed, though, because it ignores all previous activities that led to the sale, which could result in reduced or cut investment regarding the key contributing activities.

To build a truly accurate attribution model you need current and historic individual-level data from your online channels.

Read on to see why.

The value of aggregated individual-level online data

Only by piecing together each individual customer journey can you build up a truly accurate picture of the inter-relationship between your digital marketing activities and the relative role, importance and value of each one in the overall strategy.

For example; without aggregated individual-level online data, how do you know if:

• The specific PPC terms you are investing in to acquire new customers are really driving new visitors to your site - or are you only achieving already brand-aware users who would have visited your site anyway at much lower cost?
• Increasing investment in the paid search terms driving 10% of site traffic, at the expense of your display advertising, will positively impact overall results?
• You are paying your affiliates fairly based on their true contribution? You may be under-rewarding those who bring you new prospects who don’t convert on their first visit?

This example video shows how insurer, Hiscox, saved 10% of its PPC spend using full customer lifecycle attribution modelling:

Olympic TVs example: An over-reliance on ‘last click’ attribution

Take a look at an example of how the concept of digital marketing attribution modelling works, using a fictitious multi-channel television retailer called Olympic TVs.

This table shows Olympic TVs data about the last or converting click for one particular sale, and the insight it can gain from this data:

Based on this data, Olympic TVs would assign 100% of the value of the sale to Google. It might then decide to increase investment in PPC, to the detriment of its other traffic building activity, such as online display and affiliate marketing. This decision, based on an incomplete view of the role and impact of the customer’s previous interactions with Olympic TVs, might damage rather than improve results.

Olympic TVs example: Full customer lifecycle attribution

 Alternatively, if Olympic TVs considers the previous interactions of an individual with its brand over time, it gains a much more valuable picture of the role and importance of each digital marketing touch point with its brand.

The table below shows data from the individual's previous four visits, revealing some of the insight that can be gleaned from this data:

It is immediately obvious the website visit is in fact the fifth visit by this individual. Based on this additional insight, Olympic TVs now knows, amongst other things:

• Driving e-newsletter registration is an important marketing tactic as its email marketing programme drives results.
• Its affiliate programme is having an impact and, specifically, which affiliates are driving the right kind of customers.
• Even for price sensitive customers (initial search was for “cheap TVs”), quality is important as they interacted with both a product review and a campaign promoting the best picture quality.

Looking beyond the website

With the proliferation of channels and devices available to consumers nowadays, it’s important to look beyond website interactions to include, for example, the brand’s Facebook page, m-commerce sites, mobile apps and other online channels and digital technologies. The same customer might:

• Visit your m-commerce site after reading an email from you on their smartphone.
• Log onto your Facebook page on their work laptop.
• Use your mobile app on their tablet to check for offers and discounts.
• Make a purchase on their home PC.

You need to be able to connect up these multi-channel interactions to assign them to an individual.

As a result of its better understanding of cross-digital channel behaviour, Olympic TVs gains additional data and valuable insight, such as:

Even in this simple example, it’s possible to see how bringing together a rich blend of individual, current and historic multi-digital channel and device data allows Olympic TVs to start to build robust accurate digital marketing attribution models.

Olympic TVs can also attribute an actual value to each customer touch point, enabling it to make highly informed marketing investment decisions and replicate effective marketing programmes in the future.

 This data can also be used to drive segmentation, targeting, personalisation, channel optimisation and a wide variety of other marketing activities… but that’s another subject! 

Model-building considerations

Once you have all the individual-level online data, there is a series of decisions to be made in order to build attribution models that are relevant and right for your specific business; including:

•  Length of ‘look back’ period - how far back should you look to understand the influence of a customer’s interactions on the eventual sale? 
• Revenue share modelling - how should you allocate the revenue across all the customer’s interactions with you?  Beyond first and last click, other methodologies include:
• Flat - all interactions share credit evenly.
• Time decay - interactions closer to the conversion receive a higher weighting.
• Weighted - specific touch points receive increased weighting based on criteria such as which pages were visited during that interaction.
• Customised by channel - certain channels receive higher or lower weightings; for example, more weighting is given to more costly channels.
• Profitability assessment - by segmenting customers according to profitability and lifetime customer value, you can look for any differences in customer journey by customer value to ascertain whether certain activities build more or less profitable customers.

The advantages of using individual-level lifecycle data

Having the right data about individuals' interactions with your brand across all digital channels over time is an essential starting point in building digital attribution models that are capable of helping you to:

• Evaluate marketing effectiveness in terms of both channel and message/creative.
• Understand the true importance and worth of each marketing activity in driving business value.
• Prove and maximise the return on marketing investments.
• Optimise marketing spend across the digital channels.
• Allocate affiliate payments fairly.

Thanks to Katharine Hulls for sharing her advice and opinions in this post. Katharine is VP Marketing for Celebrus Technologies, a provider of software that enables you to capture individual-level data from your online channels for use across a variety of customer intelligence and marketing applications. You can follow her on Twitter or connect on LinkedIn.

With thanks to Edelman Digital for use of the image.