Our Attribution Models are one of the most important concepts to understand.

This only makes sense though -- we're an MTA solution, right?

In short, our Attribution Models assign credit for conversion events to relevant marketing touchpoints.

Northbeam offers 6 different attribution models, which are divided into two categories: Simple Attribution Models and Multi-Touch Attribution Models (MTA).

• Simple Attribution Models
  • First Touch
  • Last Touch
  • Last Non-Direct Touch
  • Linear
• Multi-Touch Attribution Models
  • Clicks-Only
  • Clicks & Views

---

Simple Attribution Models

In Simple Attribution Models, all credit is given to a single touchpoint. These are legacy models that you may be used to seeing.

Here are our simple models:

• First Touch
• Last Touch
• Last Non-Direct Touch

📘

Please Note

Last Non-Direct Touch is the default model used by most in-platform reporting (Facebook Ads, Google Ads, TikTok Ads, Pinterest Ads, Klaviyo, Attentive, etc.) and Google Analytics.

Multi-Touch Attribution Models (MTAs)

In Multi-Touch Attribution Models, credit is divided across multiple touchpoints.

Here are our multi-touch models:

• Linear
• Clicks-Only
• Clicks + Views

👍

Proprietary Models

Clicks-Only and Clicks + Views are Northbeam's proprietary models. Learn more below.

Summary of Northbeam's Attribution Models

---

First Touch

Starting off with the First Touch model. In this model, full credit is given to the first touchpoint of the customer journey.

Last Touch and Last Non-Direct Touch

Next, we have Last Touch and Last Non-Direct Touch.

In these models, full credit is given to the last touchpoint of the customer journey.

But here's the difference between the two: Last Non-Direct Touch does not give credit to Direct if it is the final touchpoint.

As an example, let's take the following customer journey that resulted in a $90 order.

This model is great to understand the impact of your lower funnel channels -- Branded Search, Organic Search, and Email/SMS to name a few.

Linear

Next, we have Linear. Linear divides credit equally across all channels.

Clicks-Only

Next, we have Clicks-Only. This proprietary model is similar to linear, where credit is divided equally between all touchpoints.

❗️

Difference between Clicks-Only and Linear

Generally speaking, Clicks-Only is like Linear, but gives an emphasis to your Upper Funnel channels.

In Clicks-Only, credit is not given to Lower Funnel touchpoints IF there are Upper Funnel touchpoints in the journey. The only instance where Lower Funnel touchpoints receive credit is if there are no Upper Funnel touchpoints in the journey.

Okay, we get it. There are a few things to unpack here. You may be asking yourself:

• "What's an Upper Funnel Touchpoint?"
• "What's a Lower Funnel Touchpoint?"
• "I don't understand the credit allocation"
• "Why isn't credit given to Lower Funnel Touchpoints?"

Don't worry - we'll be sure to cover all of these points.

📘

What are Upper Funnel and Lower Funnel channels?

Lower Funnel Channels

• Direct Visits (Labeled as "Organic" in the dashboard)
• Organic Search
• Paid Branded Search
• Email or SMS

Upper Funnel Channels

• Any channel that is NOT a lower funnel channel (ex. Paid Social, Paid Search (excluding Brand), PMax, Organic Social, Affiliate, Influencer, Display, etc)

👍

Please Note

We understand these "Upper Funnel" channels are not always used for upper funnel in your media strategy. The idea here is to differentiate Upper Funnel vs Lower Funnel to move credit around.

🚧

To better understand...

Find example customer journeys and credit allocation below.

Clicks-Only: Example 1

Let's take the following journey and assume we're on the Accrual accounting mode - where credit is given to the touchpoints.

1. Jan 1 - Facebook Ad click
2. Jan 2 - TikTok Ad click
3. Jan 3 - Google Branded Search Ad click, resulting in a $90 purchase.

Explanation: Although the Branded Search pushed the conversion, Branded Search does not get credit because this is a Lower Funnel Channel. Therefore, credit is distributed equally to Facebook and TikTok.

Clicks-Only: Example 2

Here's another example. Let's say the entire journey is made up of Lower Funnel Touchpoints:

1. Jan 1 - Google Branded Search Ad click
2. Jan 2 - Direct Visit
3. Jan 3 - Email click, resulting in a $90 purchase.

Explanation: Because all the touchpoints are Lower Funnel, credit is distributed evenly. Credit is only pulled when an Upper Funnel touchpoint is recognized in the journey.

👍

Why Clicks-Only Is Useful?

Emphasis on Upper Funnel Channels

The whole idea of Clicks-Only is to see the directional impact of your Upper Funnel channels -- as this is where budget is normally controlled and where brands scale.

Let's say there are 50 conversions attributed to a Lower Funnel channel like Branded Search. That's normally great to see, Branded Search is likely doing its job and cleaning purchase journeys. The challenge lies in its actionability. In other words, this usually isn't super actionable -- particularly from a media buying or budget adjustment standpoint. It's normally difficult to scale lower funnel channels due to its finite audience. Matter of fact, these conversions are usually a result of an Upper Funnel Channel touchpoint. Therefore, by moving the credit up the funnel, we're able to pull the curtain back and see what actually drove the conversion.

It's Conservative

Clicks-Only is strictly based on hard, click data -- like all the other models, except for Clicks and Views. What we mean by this is that we're only attributing conversions that resulted in a click from a referring channel or campaign. We're not doing any modeling, unlike in Clicks and Views. We also divide credit between the touchpoints -- resulting in an even more conservative view.

The upside here is that it's very directionally accurate. If a channel or campaign has strong performance in Click-Only, there's normally strong conviction that it's driving incremental revenue for your brand.

Clicks & Views

Your Northbeam attribution models leverage machine learning! This means your Northbeam dashboard has had the opportunity to ingest your data to stitch together customer journeys and you can start accounting for view-through data when analyzing attribution decisions.

Clicks & Views provides directionality on which ads within a channel are driving the most results and which ones are ineffective. Utilizing the Clicks & Views model in conjunction with Northbeam’s Clicks Only attribution model can help you fill in any “click-through” gaps, particularly for channels like Social and Video that are highly view-oriented.

What exactly is the Clicks & Views Attribution Model?

Clicks & Views is Northbeam’s proprietary machine learning model that accounts for impression and ad engagement data when making attribution decisions. This model looks to split attribution credit on a Multi-Touch basis, and reduces attribution credit to any direct, organic search, paid branded search, or email/SMS touchpoints in the customer's journey. The Clicks & Views model will also give appropriate credit to Views at any point throughout the customer's journey.

📘

NOTE: View-Through data measures how many visitors saw your ad but did not click.

We built this model because unattributed traffic is not helpful for marketers when trying to analyze which channels are powering their highest-converting paths. Although our Clicks Only model is our most accurate model because clicks are hard data, there are many conversions that won’t be captured -- as clicks don’t represent the entire picture.

We understand that many customers eventually convert via direct visits after watching a TikTok, YouTube, or Facebook ad. This is why our Clicks & Views model tries to assign proper credit to any ad views throughout a conversion path. This can help make your data more actionable by giving you a more complete picture of what’s going on with your campaigns.

How does Clicks & Views work?

At a high level, our model tracks engagement metrics (impressions, video views, etc.) and tries to find correlations between those metrics and any corresponding organic spikes. Our model then makes an educated guess using machine learning from your data to assign organic traffic and conversions to the ad campaign that prompted that behavior. Please note the attribution window for Clicks + Views is 1 day view only regardless of what option you select.

When do you use Clicks & Views?

Keep in mind, our Clicks & Views model takes 25-30 days to "learn". Historical data collected by our tracking pixels inform our model.

One of the most common questions we get is: Which attribution model is the best one to use to make media buying decisions? The answer is: Clicks & Views offers directionality of what views are driving incremental revenue. Look for big gaps between revenue reported in Clicks Only versus Clicks + Views to better understand what assisted the organic conversions. We recommend using the Model Comparison tool in Northbeam to do this via the hamburger icon on the top right of your Northbeam dashboard next to the maintenance alerts icon.

Let’s say you notice that Clicks + Views captures a bunch of revenue attributed to YouTube ads that weren’t captured in Clicks Only. This could mean that YouTube ads are driving revenue, but we should try and validate these results before dramatically scaling spend. One way to do that is to conduct post-purchase surveys and track where incremental revenue is coming from. You could also slowly ramp up spend on YouTube while keeping an eye on blended revenue numbers to see if they’re highly correlated. If post-purchase surveys are also indicating YouTube as a leading channel, then you can be reasonably confident that YouTube ads are having a positive and additive impact on your business.

What is the benefit of looking at the Clicks & Views model?

Ultimately, the Clicks & Views model can help you better allocate your marketing dollars. You spend less on conversion ads if more people view them, don’t click, and then later purchase.