Data & addressability | part 3 of 3 | attribution modelling
White Paper | Author: Cyprus Jake Malinao Teads, Data and Addressability subgroup, IAB SEA+India Programmatic Council
This is the final article in this 3-part series about Data & Addressability, clarifying the different types of attribution modelling and how they are best put to use.
How are we leveraging Data and Addressability in the most meaningful way to allow us to attribute our marketing spends intelligently?
As marketers, we understand that there are numerous components to a comprehensive marketing strategy. One of the most important questions we have is how effective our marketing campaigns are. There are numerous touchpoints along the buyer's journey, such as, for example, specific ads, blog posts, social media posts, emails etc. Marketing attribution modelling can provide additional insights and data. Attribution modelling done correctly can benefit CMO’s in analysing budget allocations, the different channel mix, audience trends & trending media channels.
What is attribution modelling?
Marketing attribution is the measurement technique or a method of assessing the marketing touchpoints encountered by a consumer on their path to purchase. The goal of attribution is to determine which channels and messages influenced the decision to convert or take the desired next step the most.
Attribution modelling is part of many statistical models which can be used within a digital marketing environment. In order to start with attribution modelling is to properly track the digital customer journeys, by setting the tracking properly on the media channel buys and on the event journey on the client app or site.
What are the benefits of marketing attribution?
Employing the right models will provide rich insights into how marketing ad spend is performing. This information can guide you when planning future strategies including -
optimising marketing spend
increasing ROI or ROAS
improving personalisation
improving product development
How are marketers adopting attribution modelling?
Digital marketing is still seen from a narrow perspective that prioritises branding over performance. When we look at performance, we get into the debate about the relevance of a click conversion vs. a view through conversion. The click conversion will typically have a greater impact and be more accountable to most clients. However, last click attribution is still the single touch attribution model that most marketers and clients use when using click conversion as a benchmark for their performance initiatives.
Let’s explore the limitations and potential flaws with just last click attribution modelling:
1. Limited understanding of the customer
2. Misleading metrics lead marketers astray
3. Last Touch Marketing Attribution ignores the Standard Conversion Funnel
What models are available, and which one works best?
1. Multi-Touch Attribution Modelling
Multi-touch attribution modelling is powerful because it takes into account every channel and touchpoint that a customer interacts with throughout the buyer’s journey, up until they decide to convert. It tells you which of those channels and touchpoints were most influential, and provides insights into how they worked together to influence the customer.
2. Cross-Channel Attribution modelling
Cross-channel attribution modelling is often used interchangeably with multi-touch attribution. However, their definitions differ slightly. Cross-channel attribution designates value to each marketing channel (such as paid, organic, or social media) but doesn’t look at the specific touchpoints within those channels the way that multi-touch attribution does.
3. Linear Attribution modelling
Linear attribution modelling is a type of multi-touch attribution that gives equal credit to all channels and touchpoints with which a customer interacts throughout the buyer’s journey.
4. First-Touch Attribution modelling
First touch attribution modelling gives all the credit for the conversion to the first channel or touchpoint with which the customer interacts.
5. Last-Touch Attribution modelling
Last-touch attribution modelling is the opposite of first-touch attribution modelling — it gives all the credit to the last touchpoint with which a lead interacted before converting.
6. Time-Decay Attribution modelling
Time-decay attribution modelling gives credit to all of the touchpoints that contributed to a conversion and also considers the time that each touchpoint occurred — the touchpoints that happened closest to the time of conversion are weighted most heavily.
7. U-Shaped Attribution modelling
U-shaped modelling, also known as position-based attribution modelling, splits the credit for a conversion between the first and last touchpoints.
8. W-Shaped Attribution modelling
W-shaped attribution modelling gives the most credit to the first touchpoint, last touchpoint, and mid-funnel touchpoint before a conversion — it then gives equal credit to the rest of the touchpoints.
We should all avoid the First (only) or Last Interaction (only) attribution model by today's standards. The best method for your business will be determined by your marketing needs and the complexity of your marketing mix.
If it all seems too complicated, not to worry. Start with the simpler Time Decay and Position Based models, although again, avoid First and Last Click. To make attribution modelling of all your marketing campaigns accurate, all your campaigns must also be properly tagged and easily identified so you can give credit to where credit is due.
Who can build attribution models and measurement?
Attribution modelling can be used on multiple levels and be built by different types of marketers. Client marketers can build attribution models to monitor and analyse their overall performance across multiple vendors and different media channels. Agency marketers should build attribution models for their clients to optimise building better media strategies for their clients.
Conclusion
As you can see, there are many attribution modelling tools from which you can choose. However, I believe that to be ready for the future we should be adopting attribution modelling platforms on a client level. This will provide continuous insights as the digital media landscape continues to evolve and change.
Read Part 1 on The Supply Chain of the Future by Emily Yri, PubMatic
Read Part 2 on Identity Resolution in a Cookieless World by Nisa Seah Oracle Advertising and Suhail Ahmed Dentsu Malaysia
