Solving the hardest problem in marketing - multi-channel attribution

We've become saturated in a sea of data, but many times are still so poor in actionable insights. There are numerous reasons for this, from technological and budgetary limitations to a lack of available skills, resources and understanding. Each of these pose a unique challenge to measuring and optimising marketing efforts, but there is a silver lining here - as more of the media mix becomes digital and further investment is made into upskilling marketers, media measurement will become clearer, more transparent and standardised.

Measuring paid vs owned vs earned media

This model has been a cornerstone of media measurement for many years and is still widely prevalent in the industry. However, it has become quite outdated as digital, social and mobile channels blur the line between the three. If I watch an ad on TV, go to YouTube to watch it on the brand's channel, then share it with my friends, that asset begins as paid (by virtue of being an ad), then becomes owned (when it's posted on YouTube) and earned (when it shows up in my friends' news feeds).

As a marketer, I need to understand how consumer behaviour is impacted by my marketing collateral across all these channels, whether they're offline, online, paid, owned or earned. Despite the sea of data available to us, there are so many factors we just don't know, including:

-How did the TV ad drive social behaviour and sharing?
-What was the impact of social media shares on offline sales?
-Across how many screens did these activities take place?
-How do we measure the lifetime value of customers originated from each channel?

In this mass of questions, it is easy to get confused about what problem we're really trying to solve. The first aspect of solving a complex problem involves clearly defining the problem statement, from which we can use systems thinking to break down the problem into simpler, inter-relatable parts.

© Fuzzbones – 123RF.com

A simpler question to ask may be:

"What offline sales can be attributed to online channels?"

Another question we can ask is:

"What is the online impact of our offline campaigns?"

From here, it is easier to define strategies that can be put into place to give us the right answers with a reasonable degree of confidence.

As marketing professionals, we need to keep in mind that there are no silver bullets to multi-channel attribution. We need to clearly define our measurement goals and the problems we're trying to solve, by first asking the simplest possible questions to get us to the right answers.

We also need to be cognisant of the fact that there are still massive gaps in our data, as the marketing eco-system is owned by many different players who keep data in walled silos. Despite what people may tell you, for example, the following problems have not been solved in any simple, standardised way:

It is almost impossible to track a single person's behaviour across multiple devices, unless they are logged in across all devices to a portal where the data can be accessed by the analyst. Facebook, for example, allows us to log into our PCs, phones, smart TVs, but certainly doesn't share individual usage data to marketers.

Online-only multi-channel attribution has become incredibly sophisticated, and existing analytics solutions can already track clickstream data across social, web, PPC and banners to attribute digital media spend to specific channels. However, there is no simple solution for bringing this data into the offline space, e.g. cookie data cannot easily be linked back to an offline credit card purchase.

With online remarketing it has become easy to target people across different platforms based on their past behaviour, but it is still not possible to link cookie data back to individuals on social media. If I click on a promoted tweet for an e-commerce campaign but abandon my cart, a social community manager can't message me to start a conversation about my pending purchase. This is obviously for good reason, as the privacy implications would be disastrous.

With all this in mind, I'd like to present a few strategies that can be put in place to answer the previous questions put forth.

"What offline sales can be attributed to online channels?"

Strategy 1 - Loyalty programmes

With a well-implemented loyalty programme, it becomes easier to track an individual across different channels and link them to a loyalty account or club card. Many local retailers use this to great effect. Customers can sign up on a website to receive a loyalty number, which they then use in-store afterwards.

Strategy 2 - Customer surveys

The tried and tested strategy of simply asking your customers about the product special or campaign works amazingly well. Plenty more data can be collected if this is linked to a reward such as a competition entry or product discount. It is vital to ask the right questions here to ensure the insights are not skewed.

Strategy 3 - Promotional codes

In-store promo codes and coupons can be used to track customers who have claimed them online. Linking the online coupon codes to the offline purchases creates a direct data link between online exposure and offline behaviour, which can result in a mass of insights. Its important to generate coupon codes that can be traced to individual online media channels, days, times and even specific ads. The coupon then becomes an aggregated hash of the consumer's online exposure..

Strategy 4 - Unique CTAs

Unique call to actions that are not marketed offline can be used to gauge behaviour from an online campaign. An example of this is a telephonic call to action with a number that is unique to the online campaign. Another example would be directing people to an activation at a specific time or place, and tracking the correlation between this behaviour and online exposures. Although not perfect, these kinds of creative measurements can go a long way towards providing cross-channel insights.

"What is the online impact of our offline campaigns?"

Strategy 1 - Unique URLs and hashtags

Many local brands include unique URLs or hashtags in their above-the-line campaigns. A unique shortened URL can be easily shared on radio or a billboard, which can be tracked through online analytics solutions. Hashtags on billboard, print and TV ads can be used to stimulate online conversation and can be tracked using social analytics tools, but still need to be used in a creative way to maximise engagement.

Strategy 2 - Online surveys

Similar to the offline strategy, an online survey is incredibly quick and easy to put together, and can be included as part of the consumer's purchase path as an easy way to aggregate data. For companies with a large number of respondents, this data can be segmented to assist with campaign strategy in the future.

Strategy 3 - Data correlation

Although attribution is all about finding causations between data, it is much easier to measure correlations in siloed data sets. Correlations can be misleading, simply because they can be based on data that has no connection, or misunderstood third-party factors can influence both datasets simultaneously. Just because the number of pirates in the world has decreased while global temperatures have increased, does not mean that the lack of pirates is causing global warming. In campaigns, hourly and daily media-spend can be correlated with social media shares and engagements, website traffic, and online purchases to find any potential links in the data.

There are many other strategies that can be implemented for specific campaigns, but it is important to note that there is no one-size-fits all solution to measurement. We must always tie our insights back into our business goals and problem statements so as to understand what we're measuring, which will inform our measurement strategy going forward.