Having explained the macro-environment related to attribution in part 1, we will be discussing all factors affecting the accuracy of data collection in this article, in order to guarantee a good foundation for attribution.
As a recap, attribution is meant to determine the effectiveness of your marketing activities and optimise performance. In other words, e-commerce can identify the strategic marketing activities that bring in more sales, as well as better allocate resources. For that, they turn to technologies for help.
However, it isn’t easy setting apart nice looking analytics tools from the ones with accurate data. Yet it is fundamental to do so. After all, what good is attribution when the data behind it is false? That is the equivalent of saying “if my data were to show…” without studying what is presently true – purely theoretical.
And the accuracy of your data is determined by the definition of two key processes: setting the right data structure and the actual collection of data.
Setting the data structure straight
To avoid losing sight of the big picture, there needs to be a framework for all interactions that consumers have with a brand or product. A common framework for the rough sorting of marketing activities is the distinction in paid, owned and earned channels, where:
- Paid refers to all paid marketing activities in partnership with external providers for reach and access to potential consumers, such as search engine advertising, social media advertising, and through affiliate partners
- Owned refers to your own marketing activities, such as emails to existing customers or a self-managed community – as a rule of thumb, costs for owned media are incurred only for the technical implementation of the marketing measures
- Earned are all contacts that are generated free of charge and often because of other marketing activities, such as direct access to the website, links from other websites, organic search engine results, and social media interactions
This means that e-commerce, who intend to conduct attribution to draw insights on their marketing strategy in its entirety, must ensure that all of these interactions can be captured. At the very least, if the goal of the attribution is the optimal distribution of marketing budgets, all paid interactions should be recorded.
Understanding the different data collection methods
Most e-commerce today gather data using tools meant for tracking, container management and analysis. Often, many of these solutions serve media too, and thereby tend to deliver data favouring their own channels or media. Since the quality of attribution rules depends on the accuracy and depth of the data collected, any non-neutral information will alter the actual results at the end.
Which data collection methods does your technology use?
For paid channels, redirects are standard to ensure 100% accurate allocation of traffic.
A major drawback, however, is the resulting bottleneck, as a significant proportion of marketing activities is managed via one central infrastructure; Any failure will cause important sales information to be lost.
- URL parameters
URL parameters can be added to redirects to feed other systems used in parallel with appropriate data. However, this can easily fail due to technical features of third party tools, causing traffic to be incorrectly allocated, e.g. to direct type-in or organic links.
Referrers contain the previously visited website. They are highly unreliable, as they can be manipulated with little technical effort. It often becomes the pool for unassigned or non-tracked traffic from other methods.
As each collection method has its pros and cons, it is important to ensure your technology is neutral and supports a combination of them for the best possible data collection, and sorted into the defined data structure of the e-commerce’s choice.
Factors affecting the accuracy of data collection
While the methods of collecting data play a large role in the accuracy of the data collected, there are many other obstacles in piecing together the exact customer journeys.
When it comes to tracking, the goal is to ensure that all possible interactions a customer has with your brand or product are tracked. However, there are multiple ways of hindering tracking, such as the use of ad-blocking software, or disabling tracking via browser and device settings. These are often the result of customers hindering tracking on purpose as they no longer wish to be impacted by poorly targeted ads. On the e-commerce’s side, there may also be scripts or website errors that have not been solved, causing tracking to be unsuccessful.
One of the best ways to minimize these hindrances is to use first-party cookies, implemented directly via your domain. Also, there are other tracking methods that do not involve cookies. Instead, the technology matches technical information from the user’s browser and device to patch together multiple touchpoints into one customer journey.
Disruption through change of media
In today’s omni-channel world, consumers are constantly switching between channels and devices. Often, this causes tracking breakpoints, which translates to a loss of or misinterpretation of customer journeys. Here are the most common disruptions and little workarounds for them:
- Offline to online
E-commerce could overcome this by using separate domains, dedicated landing pages or vouchers to connect the journey from having interacted the brand offline, to accessing information online.
- TV to online
This could be overcome via TV tracking or similar measures as offline to online.
- Desktop to mobile web
This is one of the most common breakpoint. The best way is to encourage users to log in to their profiles both on desktop and mobile websites, or to engage Data Management Platforms (DMP) to match user IDs to identified users.
- Mobile web to app
Mobile tracking SDKs like Adjust can help beat this breakpoint.
- Email to browser
E-commerce can activate impression tracking, so there is a touchpoint on the customer journey that identifies that the interaction prior to the browser is based on the viewing of an ad media on an email.
Generally, advertising platforms are open and generous with their collected data – this is for as long as they are not in a monopolistic position. The larger the market player, however, the more closed the ecosystem. They usually present data that is sampled or aggregated, with incompatible technical standards, making it difficult to create a connection with other third-party systems.
The Walled Gardens (e.g. Google) are cushioned by advertising giants (e.g. AdWords), so you feel at home working with them. Their wide reach make them unavoidable, and they offer attribution and conversion tracking based on rules that are neither neutral nor budget-independent.
The workaround for e-commerce is to share data and services with the big monopolists as much as necessary, but also as little as possible. Therefore, by using an independent tracking and attribution solution, AdWords, Criteo and Facebook are decoupled from their rules. This way, you can create cost-to-sales ratios according to your rules.
With all the data collected, we will find out in the next instalment how much of your data collected can be used for meaningful attribution.
Here is part 2 of the masterclass by COO Robert Schneider. The masterclass was held in German.