Chord Commerce Data Platform
...
Data Models
Marketing Attribution

Multi-Touch Attribution Models

8min



Need help getting started? Check out our marketing attribution Quickstarts for more information!

Understanding Attribution Models: Enhancing Marketing Performance

Attribution models are pivotal in enhancing marketing performance by allocating traffic and sales to specific marketing channels. These models offer valuable insights that enable businesses to gauge the impact of various marketing efforts accurately.

In the past, Chord exclusively supported first-touch attribution at the visitor level, which helped clients identify the channels responsible for attracting visitors and generating initial website visits and purchases.

However, Chord has now introduced a groundbreaking feature - multi-touch attribution models. This advancement includes three new types of attribution methodologies, allowing businesses to better understand their marketing strategies' effectiveness. With these enhanced models, marketers can make more informed decisions and optimize their marketing campaigns for better results.

Why use multi-touch Attribution Models

Multi-Touch Attribution Models are essential because they provide a more comprehensive and accurate understanding of the customer journey and the contribution of each marketing touchpoint to conversions and sales. They enable your business to gain deeper insights into your customer behavior, optimize marketing efforts, and make informed decisions, leading to improved overall performance and higher ROI.

  • Realistic representation of customer behavior: Multi-touch attribution considers all touchpoints, reflecting how customers interact with multiple channels before making a purchase.
  • Better credit allocation: Multi-touch attribution models distribute credit across touchpoints, identifying critical channels influencing conversions and sales.
  • Optimization of marketing spend: Identifying influential touchpoints helps allocate resources efficiently, maximizing ROI.
  • Insights into customer behavior: Multi-touch attribution provides valuable insights into customer interactions at different journey stages, enabling targeted campaigns and improved conversions.
  • Enhanced decision-making: A holistic view empowers data-driven decisions, comparing marketing strategies and refining approaches.
  • Improved team collaboration: Multi-touch attribution aligns departments, highlighting contributions to the overall customer experience and encouraging a customer-centric approach.

Only sessions before a user's first purchase are counted toward attribution.

What is User Stitching?

User stitching refers to associating a single user ID with all recorded events related to a particular user. While this may seem straightforward, it becomes challenging in a digital landscape where users interact with websites using different devices and across multiple sessions before and after making a purchase.

Although user stitching is not mandatory for marketing attribution, it offers distinct advantages. While sessions can be partitioned using anonymous IDs, attributing events to specific user IDs provides more precise and accurate attribution. As a result, user stitching is preferred to ensure a more comprehensive understanding of user behavior and improve marketing attribution efforts' accuracy.



Channel Mapping: Assigning Sessions to Refer Channels

In channel mapping, we assigned a referring channel category to each session. This involved extracting sources and mediums from UTM tags attached to the sessions. In cases where sessions lacked UTM tags, we inferred the source and medium from the referring URL when possible.

By mapping source-medium pairs with specific channels, we successfully allocated all sessions into distinct categories, including:

  1. Paid search
  2. Organic search
  3. Paid social
  4. Organic social
  5. Email
  6. SMS
  7. Direct
  8. Affiliate
  9. Other

These high-level categories provide an overview of session attribution. However, we can delve deeper into attribution analysis using UTM source and UTM medium for a more granular understanding to uncover more detailed insights.

How is attribution determined in Chord's data models?

Our current Attribution Rules:

  • SEM: Campaign source contains a search engine domain name and campaign medium equals cpc or cpm.
  • Paid Social: Campaign source equals a social domain, and campaign medium equals cpm.
  • SEO: The campaign source is empty, and the referrer contains a search engine domain name.
  • Affiliate: Campaign source equals pre-defined affiliate campaign names.
  • Organic Social: Campaign source is empty, and the referrer contains a social domain.
  • Employee: Referrer contains ‘staging.’ or unique_id contains (‘%chord.co’)
  • Influencer: Campaign medium equals ‘influencer.’
  • Ambassador: Campaign source equals ambassador.
  • Referral: Search contains ‘?ref=’
  • Organic: Referrer is empty: All other visitors that don’t fit into the above-described categories. So with these definitions, direct traffic currently falls in Organic.
  • Other: if a session does not have a UTM source,does not have a UTM medium,does not have a Gclid, these sessions will be automatically mapped to the Other channel

In a perfect world, Chord has access to a UTM source, medium AND a referrer host for every session that passes through our models. Unfortunately, this is not realistic. Because, in many cases, some sessions will only have a combination of these 3 attributes (i.e.utm source, utm medium, referrer host), sessions can be manually mapped on the Chord data side so that they automatically are binned into the channel that you expect.

For any session where there is at least 1 present attribute (it can be UTM source, UTM medium OR referrer), this session can be manually mapped to whichever channel your team prefers. Additionally, if you're a Shopify customer, you will see First appended to the marketing attribution fields.

How to build a multi-touch Attribution Model Report

Check out our marketing attribution quickstart to start building your attribution report!