How Conversion Modeling Works for Ecommerce

In early 2024, Google updated how it collects data, in order to comply with new privacy laws in the UK and Europe. Under the new framework, websites must now obtain user consent to track visitors. If the user declines, companies like Google (and, by extension, ecommerce merchants who rely on its services) lose access to key user behavior insights. This includes repeat visits, referral link usage, or whether a customer journey ends in a conversion.

To address this data gap for businesses, Google introduced conversion modeling, a machine learning tool predicting user behavior and estimating modeled conversions when direct observation isn’t possible. Conversion modeling offers key insights into your user base, helping you learn what drives them to your ecommerce site and identify which areas you may need to invest in more. 

Here’s how to use conversion modeling to boost your business.

What is conversion modeling?

Conversion modeling is a data-driven marketing approach using machine learning and other statistical methods to predict and optimize user behavior—especially when direct tracking is unavailable. The goal is to estimate conversions—customer actions like purchases and sign-ups—by analyzing aggregated or anonymized data patterns.

Websites collecting cookies are like a dinner party distributing coat check tickets; unique IDs connect each guest to their “coat” throughout their time browsing. New laws, however, allow guests to decline a ticket (reject data collection), making it harder to track them and their coats (i.e., their information). Thankfully, by using available conversion data from countless past parties, you can predict which coat belongs to whom and map out their journey.

How does conversion modeling work?

1. Separate ad interactions
2. Subdivide the two groups
3. Link unobserved and observed interactions

Conversion modeling may seem complicated and overwhelming, but it can be broken down into a simple three-step process. Here’s a quick explanation of how the framework functions:

1. Separate ad interactions

For search engines and ad companies like Google, the first step of conversion modeling is to divide users’ interactions with ads into two groups: observed and unobserved. 

• Observed. Observed actions include ad interactions trackable via cookies. These have a clear connection to a conversion. This data includes page views, user counts, events, and sessions.

• Unobserved. The unobserved group comprises users who didn’t consent to tracking and therefore have less available data.

2. Subdivide the two groups

The observed and unobserved ad interactions are then split into subgroups based on shared characteristics—like general user location or time of day—you can still collect regardless of user preference. You’ll then calculate important metrics like conversion rate for the observed groups. For example, you may find user conversion rates are higher in the southern US on weekday evenings than on weekend mornings.

3. Link unobserved and observed interactions

Observable data points with shared characteristics get linked to their unobserved counterparts to predict unobserved conversions. For example, using the information about observed users in the southern US, the framework infers that unobserved users in the same category may behave similarly (modeled data). Machine learning models further enhance these predictions by incorporating extensive third-party training data.

How conversion modeling can benefit your business

• Gain accurate advertising insights
• Add a competitive advantage
• Build customer trust

While still relatively new, conversion modeling has already proven its value in getting the most out of companies’ marketing and advertising efforts. Google claims conversion modeling can recover up to 70% of lost pathways between ad interactions and conversions caused by user consent choices. Here’s how conversion modeling can benefit your business:

Gain accurate advertising insights

With conversion modeling in place to bridge gaps in understanding customer behavior, you can learn how something like a Facebook ad campaign drove potential customers to your site. It can also show how paid Instagram ads are leading to conversions—increasing your marketing return on investment (ROI).

Add a competitive advantage

In the fierce digital marketplace, even a minor insight into customer behavior can differentiate your business from competitors and elevate your sales strategy. Conversion modeling, as a relatively recent development, can provide a boost. 

Many companies struggle to adapt to the new digital landscape and continue to rely on traditional tracking methods. However, employing conversion modeling can provide you with unique data by filling in the information gaps of untracked users, giving you an advantage over your competitors.

Build customer trust

In recent years, public anxiety about data collection has surged, with 46% of individuals surveyed in 2023 expressing concern over companies tracking their information.

The new privacy rules aim to empower consumers by giving them control over data collection. Complying with these rules and respecting user consent choices allows your business to demonstrate transparency. It’s an opportunity to show customers you genuinely value their privacy concerns by offering a thoughtfully written consent pop-up on your site. You can install pop-ups on your Shopify store by installing a Shopify app.

Best practices for conversion modeling

• Define your goals
• Work with an expert
• Monitor and analyze data
• Choose the correct offering and test it
• Be patient

Getting the most out of your customer data tracking with conversion modeling requires planning and investment. Here are a few tips and tricks for taking advantage of conversion modeling in your business:

Define your goals

Why do you want to collect user data, and how do you plan to use it? Is your goal to streamline your paid search campaign, boost customer retention, drive repeat visitors, or increase sales? Establishing clear objectives informs how you implement and leverage conversion modeling.

Work with an expert

Conversion modeling is a complex, recent framework. Rather than try to figure it out entirely on your own, rely on a qualified expert. Many digital marketing professionals understand conversion modeling and can help you with its implementation, troubleshooting, and fine-tuning to align with your ecommerce goals.

Monitor and analyze data

Once your site has conversion modeling in place, use the behavioral data to better inform your marketing and advertising strategies. Tools like Google Ads conversion modeling can generate reports integrating metrics like user behavior, historical trends, device type, and conversion paths, offering a comprehensive view of the customer journey.

Choose the right offering and test it

While Google Analytics offers perhaps the most well-known pairing with conversion modeling, other analytics platforms may be a better fit for your business. Consider the following:

• Heap Analytics. Heap offers automated insights into user data out-of-the-box, as well as a replay-style tool. It allows your business to go over any customer journey in detail and determine pain points or areas of success. Its free plan includes up to 10,000 monthly sessions. 

• Matomo Analytics. An alternative to Google Analytics, Matomo allows businesses to own their analytics data, as well as import any historical information collected from Google. Plans start at $26 per month.

• Snowplow. Snowplow offers granular, real-time behavioral data collection and processing, as well as extensive customization allowing your business to unlock actionable insights. Snowplow has a free open-source option, with custom pricing for pre-built plans.

After selecting a platform, take advantage of trial periods to ensure it meets your business needs and supports your goals.

Be patient

Gathering enough aggregated data and historical conversion rates to generate a meaningful report on your marketing and advertising efforts takes time. Even then, shifts in strategy may be incremental. Allow your business time to collect data to assess the impact of new digital marketing or investments in marketing channels like paid search campaigns.