Data-Driven Attribution and How it Differs across Google Products

According to Gartner, about 74% of CMOs expect to spend more on digital advertising in 2021 than they did in 2020. But how can you assess your channels to know exactly where to invest more? Which ads make potential customers move to the next step of the funnel?

The solution is hidden in attribution — how the value of a conversion is distributed across channels that move the user through the funnel. However, some attribution models show you only part of the picture. And these gaps in data might be critical. After all, according to the rule of seven touches, the actual purchase frequently happens only at a customer’s eighth interaction with a brand. However, all steps affect one another and eventually lead to the conversion. So how can we objectively assess the conversion path?

As an ad giant, Google offers multiple attribution solutions, from standard attribution models to advanced options with the possibility to track multiple channels. In particular, several products allow you to set up a Data-Driven Attribution model that will help you dive deep and accurately credit marketing channels.

But how can you decide which service will best fit your business? What’s the difference between Google Ads and Search Ads 360? In this article, we review and compare the most popular products that offer Data-Driven Attribution.

Data-Driven Attribution (DDA) is an algorithmic model that allows you to analyze all relevant data and assign the right credit to every customer touchpoint.

Data-Driven Attribution by Google focuses on your advertising account’s data as a unique starting point for analysis. Unlike standard models with predefined formulas, DDA uses algorithms to analyze every case differently and assess the mutual influence of channels in the funnel, even if it is complicated, inconsistent, and multi-step.

The data-driven attribution model is there to satisfy the evolving needs of marketers and provide more accurate results compared to previous-generation models. It’s used to get a better understanding of what works and what was a bad idea, as well as to improve the general performance of the marketing team.

To satisfy various business needs, Google offers DDA in a range of services. The differences between these services are in the data analyzed, the algorithms applied, and the level of customization. Some of them are designed only to track ad clicks and optimize keywords and paid campaigns, whereas others provide a full analysis of a customer’s online journey.

Before selecting a particular product, consider the following:

• What’s your advertising budget?
• What are your business goals?
• How many conversions do you have on average every month?

Now, let’s take a closer look at what products Google offers that include Data-Driven Attribution.

Google Analytics 4 continues to develop and is making features previously available only to large corporations free and accessible to everyone. On January 7, Google released a revolutionary update to Google Analytics: now, the cross-channel data-driven attribution model (DDA) is available to all users.

In Universal Analytics, single-channel attribution models were applied by default. These models included Last Non-Direct Click, First Interaction, Last Interaction, Linear, Time Decay, and Position Based. Here’s an example of how conversion value is distributed in single-channel attribution models:

Users of Google Analytics 360 also had access to the data-driven attribution model. However, there were additional requirements for using it, such as having a Google Ads account with at least 15,000 clicks on Google Search and a conversion action with at least 600 conversions within 30 days.

Unlike in previous versions, in Google Analytics 4, it’s possible to change the preset attribution model for all reports. Now you can apply not only cross-channel rule-based models but also a data-driven alternative for free:

• Data-driven attribution
• Cross-channel rules-based attribution (last click, first click, linear, position based, time decay)
• Ads-preferred last click attribution

The GA 4 support page states that each data-driven model is specific to each advertiser and each conversion event and explains how data-driven attribution works:

“Data-driven attribution distributes credit for the conversion based on data for each conversion event. It’s different from the other models because it uses your account’s data to calculate the actual contribution of each click interaction.”

As Krista Seiden (Analytics Evangelist, Founder & Principal Consultant at KS Digital) states, in Google Analytics 4, the number of touchpoints used in modeling goes up to 50+, ensuring that none of your marketing efforts are missed while calculating and assigning credit.

Note: As conversions can happen days after an interaction with ads, the lookback window settings go hand in hand with attribution settings. In Google Analytics 4, there are two options:

1. Acquisition conversion events (first_open and first_visit). The default lookback window is 30 days. It can be changed to 7 days if needed.
2. All other conversion events. The default lookback window is 90 days. It can be changed to 30 days or 60 days.

Changes to the lookback window apply to all reports within your Analytics property.

The attribution model you use matters a lot, and there’s no one correct choice for all businesses. You have to experiment and find what’s best for your company. However, the ability to use data-driven attribution for free is a great opportunity you shouldn’t miss.

Pros of DDA in Google Analytics 4:

• The ability to more accurately assign credit across multiple channels
• The ability to demonstrate the impact of each channel on the ultimate conversion
• Flexibility, as you can keep changing the model and reports will be adjusted accordingly

Cons of DDA in Google Analytics 4:

• It’s a black box. There’s no explanation of how exactly the modeling is being done. All you have is your faith that Google Analytics knows better.
• The algorithm must receive enough data to show the most accurate results.

Despite the uncertainty of how data-driven attribution works, Google recommends using it, since no other rules-based model can provide you with an understanding of the impact of multiple touchpoints on customer conversions. That’s why, for now, it’s the best model in Google Analytics 4.

With Google Analytics 360, you can use Multi-Channel Funnels (MCF) Data-Driven Attribution based on the Shapley Value method. This algorithm analyzes the path of your users through existing touchpoints, then creates an alternative variant where one of the touchpoints is missing. This shows you exactly how a specific channel influences the probability of a conversion. 

Data-Driven Attribution assesses data from organic search, direct traffic, and referral traffic along with all the data that you’ve imported to Google Analytics, including data from other Google products (e. g. Google Ads, Campaign Manager 360). With DDA in Google Analytics 360, you get an overview of all users’ online actions in your funnel and how each channel influences conversions. This option is most suitable for large websites with high volume of conversions.

Let’s check Google Analytics 360’s minimum requirements for using DDA along with the pros and cons of DDA in this tool.

Minimum requirements:

• A Google Ads account with 15,000 clicks and 600 conversions during the past 30 days
• Ecommerce Tracking or Goals must be set up

If you meet these requirements, you can start using DDA in Google Analytics 360. To keep using it, you have to meet the following minimum conversion threshold for the past 28 days:

• 400 conversions of each type with a path length of at least two interactions
• 10,000 interaction paths in a specific view

Pros of DDA in Google Analytics 360:

• Get a full analysis of a customer’s online journey
• See which ads, keywords, and campaigns have the biggest impact on conversions
• Distribute credit for revenue based on past data for a conversion
• The amount of credit assigned to each touchpoint depends on order of touchpoints
• Data analysis starts immediately, and the report on your first model becomes available within 7 days

Cons of DDA in Google Analytics 360:

• High cost of an account: starts at $150,000/year
• Hidden calculation logic: no explanation in the report
• Requires a consistently high number of clicks and conversions
• Doesn’t include offline data (phone calls, transactions in CRM)
• Requires a Google Ads account

The default attribution model in Google Ads is last click, but if you meet the minimum requirements you can configure Data-Driven Attribution. By default, data-driven attribution analyzes all clicks on your ads but not the entire customer journey. Based on these clicks, the model compares users who purchase to those who don’t and identifies patterns among those ad interactions that lead to conversions. To increase the number of conversions, you can use an automated bidding strategy that’s optimized based on information from the DDA model.

In contrast to Search Ads 360, Google Ads doesn’t allow you to run marketing campaigns across multiple engines and provides less detailed reports.

This product is suitable for medium-sized and bigger businesses that need to optimize marketing campaigns and keywords.

Now, let’s get to the minimum requirements and compare the advantages and disadvantages of using DDA in Google Ads.

Minimum requirements:

• 3,000 ad interactions in supported networks in the past 30 days
• 300 conversions in the past 30 days

To continue using this model, you have to meet the following minimum conversion threshold for the past 30 days:

• 2,000 ad interactions
• 200 conversions

Pros of the DDA model in Google Ads:

• Helps you optimize keywords and paid campaigns
• Helps you optimize bidding
• Shows which ads play the most important role in reaching your business goals

Cons of the DDA model in Google Ads:

• Don’t get the entire overview of the online user journey
• Need to maintain the necessary level of conversions and clicks for 30 consecutive days before you can see data in Google Ads
• If your data drops below the required minimum, the attribution model will automatically be switched to Linear

Search Ads 360 helps you manage marketing campaigns across multiple engines (Google Ads, Microsoft Advertising, Yahoo! Japan Sponsored Products, Baidu, and Yahoo! Gemini) due to native integration with the Google Marketing Platform.

By default, Search Ads 360 uses the last click attribution model, but you can also configure DDA if you meet the minimum click and conversion requirements. Unlike Google Analytics 360 and Google Ads, Data-Driven Attribution in Search Ads 360 analyzes activities in Floodlight, the conversion tracking system for the Google Marketing Platform. The attribution focuses on paid marketing campaigns and shows you how clicks on keywords influence conversions. You can also adjust or create a new bid strategy that will automatically optimize bids based on the model’s data.

The Search Ads 360 service is suitable for websites with a high number of conversions who need to optimize their paid campaigns.

Let’s see the minimum requirements for and the pros and cons of using data-driven attribution with Search Ads 360.

Minimum requirements:

• 15,000 clicks in the last 30 days
• 600 conversions in the last 30 days

Pros of using DDA in Search Ads 360:

• Get reporting data in near real time
• Optimize bids automatically using Smart Bidding technology together with DDA
• Create up to five DDA models to compare data with different channel groupings
• Possible to upload offline conversions
• Accounts for cross-environment conversions

Cons of using DDA in Search Ads 360:

• Ignores search and display impressions
• Might be not fully accurate: Search Ads 360 uses machine learning and historical data to model the number of conversions if it’s not possible to measure all conversions
• Only tracks the number of conversions attributed to paid search
• Additional setup required to realize all advantages: Campaign Manager, a set of Floodlight activities, and Search Ads 360 Natural Search reporting
• Impossible to analyze conversions tracked by Google Ads, Google Analytics, or other conversion tracking systems

Google’s Data-Driven Attribution model is one algorithmic model that can ensure a granular approach to analyzing your data. Just like data-driven attribution by Google, OWOX BI ML Funnel Based Attribution assesses the effectiveness of your advertising campaigns and channels on the customer’s way through the funnel. It also provides you with real-time reports and allows you to import calculations to optimize bids. 

Unlike the Google model, however, OWOX BI attribution is based on Markov chains — a sequence of events in which each subsequent event depends on the previous. Using this algorithm, OWOX BI attribution shows how difficult it is to move from one step to another: the higher the difficulty of moving on from a step, the greater the value a channel receives.

On top of that, due to transparent calculations, you get a solid understanding of the figures behind each report so you can safely reallocate your budget. Finally, in comparison with Google products, attribution by OWOX BI provides meaningful results with smaller amounts of data required for analysis.

Let’s take a look at what you get with OWOX BI attribution.

Minimum requirements

The minimum number of conversions depends on the number of sessions. For objective results, we recommend the following correlation between sessions and conversions:

Pros of OWOX BI ML Funnel Based Attribution:

• Track a user’s offline actions
• Control purchases and returns in your CRM
• Assess the effectiveness of each advertising channel
• Customize your funnel according to your business needs
• Exclude unmanaged channels from your assessment
• Compare funnel stages and evaluate their effectiveness
• Analyze data based on thousands of projects with machine learning
• Figure out a specific approach for each user cohort
• Get ready-made reports in OWOX BI Smart Data
• Use gathered data to manage bids and audiences

Depending on the way your business works, your customers may interact differently with your advertising campaigns. Consequently, you may need to apply various attribution models to get the best insights into your performance and how you can optimize your customer journey.

We’ve put together lots of information that will help you compare and apply modern attribution models to take your company’s marketing to the next level:

• How to choose and test an attribution model
• What is a marketing attribution model for marketers: the definitive guide
• Why your marketing needs to be data-driven
• Why single-channel attribution models are a dead end for CMOs

There are two ways to apply a new attribution model to your data.

1. If you have the Editor role on the account, you can make changes in the Attribution Settings panel under the Administrator / Property column. Here, you can select a data-driven attribution model or any other cross-channel rules-based model, or you can choose the Ads-preferred last click model from the drop-down menu.

Note: The Admin Attribution Settings do not impact attribution models selected in the Attribution reports.

2. Any user with the Viewer role can select a data-driven attribution model in the Conversion Paths and Model Comparison attribution reports.

Note: Attribution models were introduced on different dates. This means if you select a date range that includes a timeframe before the start date for the model, you will see partial data.

• Cross-channel rules-based models: June 14, 2021
• Cross-channel data-driven attribution: November 1, 2021

Changes to the attribution model will apply to both historical and future data over all reports that use event-scoped traffic dimensions. For example, you’ll notice changes to such metrics as Conversions, Total revenue, Purchase revenue, and Total ad revenue. Also, note that you can apply changes to this setting as many times as you want. In case you need more information about attribution settings in Google Analytics 4, check the Analytics Help.

We hope you have already installed a new version of Google Analytics and have begun to collect data with its new approach. Don't miss out on new features; book a call and get answers to all your questions! If you’re still unsure about moving to the latest version and saving your data, OWOX BI specialists will help you.

As the industry continues to change fast, every marketing dollar counts, and arriving at a data-driven attribution model means better performance. With a changing marketing landscape, data-driven attribution is a great tool for marketers to unlock new insights into consumer behavior and meet their goals.

Google products that offer Data-Driven Attribution allow you to track different channels, determine which online ad is the most and least effective in Google Search, and analyze users’ online journeys in detail. Even though Data-Driven Attribution by Google is generally considered as one model, its implementation differs across products. To effectively measure data, you need to choose a service that fits your data type. Here are the primary focuses of each product:

• Google Analytics 360 / Google Analytics 4 tracks all user actions, clicks, and displays based on multiple channels and their interrelations in the funnel.
• Google Ads tracks ad clicks in Google Search.
• Search Ads 360 tracks Floodlight activities and paid campaigns.

With OWOX BI, you don’t have to select among several services. You can get the benefits of Data-Driven Attribution by Google with transparent calculations on top and fewer minimum requirements all in one product.