L'Oréal Luxe achieved 57% sales growth for Lancôme's D2C website. How? Read case study to find out
All articles
Why Conversion Modelling is the easiest way to get better results from Smart Bidding ad campaigns

Why Conversion Modelling is the easiest way to get better results from Smart Bidding ad campaigns

Why Conversion Modelling is the easiest way to get better results from Smart Bidding ad campaigns Pavel Petrinich
Why Conversion Modelling is the easiest way to get better results from Smart Bidding ad campaigns
3 min read

Marketers rely on Google Ads and Facebook Ads to drive conversions (purchases, leads, or bookings) for their business.

Both platforms now use Smart Bidding (or Automated Bidding) — ML-based technology that manages the advertiser’s budget in order to drive as many conversions as possible.

How does smart bidding work

In order to maximise Smart Bidding performance, it’s crucial to provide both Facebook Ads and Google Ads with the information about the value of each click as soon as possible.

A lack of value signals or a delay in such feedback prevents ad platforms from efficient learning, targeting, and ad campaign optimisation. All resulting in a massive loss in potential revenue.

Conversion tracking and attribution are broken

While both Google Ads and Facebook Ads drive hundreds of visitors to your website every day, only a small portion of them convert within a single visit. It means that almost all paid clicks do not get any value, and Smart Bidding algorithms are not receiving enough signals to learn & optimise campaigns efficiently.

But is this how it’s supposed to be?

Does this really mean that most of your ad spend is worthless?

Of course not.

In truth, the vast majority of people who visited and did not convert right away will simply convert later, often from a different traffic source, browser, or device. However, long and complex customer journeys, paired with modern-day cookie restrictions, result in ad platforms being unable to track conversions properly.

This is particularly problematic for businesses that promote higher-value products or services (both B2B and B2C). Due to the nature of their business, they might struggle to get even the minimum amount of conversions required for Smart Bidding to work, which is typically called the “Limited Learning” issue.

As a result, up to 90% of clicks don’t get the credit they deserve, which prevents ad platforms such as Google Ads or Facebook Ads from efficient targeting, bidding, and optimisation.

But what if you could evaluate each website session immediately and provide ad platforms with valuable insights about the value of each ad click? This would give a much higher number of feedback signals for Smart Bidding to learn, which in turn should drive more conversions from the same budget.

This is exactly why we’ve built SegmentStream — a Conversion Modelling Platform.

What is Conversion Modelling and how does it work?

Conversion Modelling Platform — is a new solution by SegmentStream that helps digital marketers achieve better results from their Google Ads & Facebook Ads campaigns, by assigning a proper value to each paid click, immediately.

To achieve this, SegmentStream’s platform evaluates each website visit and with a very high accuracy predicts the Probability-to-Convert for each user during the nearest future.

It is possible thanks to an advanced machine learning model that analyses hundreds of behavioural data points on the client’s website and finds correlations between real-time user behaviour and their intent to purchase.

As soon as a visitor leaves the website without converting, SegmentStream creates a ‘Modelled Conversion’ and attributes a value to it based on the user’s probability to convert.

How SegmentStream's Conversion Modelling Works

For example, if the average conversion value is $100, and the probability to convert for a user during the session was 10%, the paid click will get the value of $10.

This way, Conversion Modelling guarantees that each ad click gets the value it deserves, regardless of the user journey and the capabilities (or lack thereof) of attribution platforms to track it. No longer your ad clicks will receive zero value if the user is simply converted from another device, browser, or cookie.

How to use Modelled Conversions to get better results from Facebook & Google Ads?

Let’s take as an example an e-commerce shop that gets 80 conversions for every 10,000 clicks. In this case, ad platforms would only know that just 0.8% of all paid clicks were valuable. The remaining 9,920 clicks would not get any value, and ad platforms would struggle to learn and optimise effectively.

Smart Bidding Learning and Optimisation without Conversion Modeliing

With SegmentStream, both Facebook Ads and Google Ads will get a much higher number of feedback signals about the true value of each ad click. Moreover, there will be no delay in the feedback because a Modelled Conversion is sent back to ad platforms almost immediately after the end of each single website session. This guarantees that each ad click gets the value it deserves.

Smart Bidding Learning and Optimisation with Conversion Modeliing

Of course, not all non-converting visits are valuable. While some portion of them will still get zero or almost zero value, yet still, there is a majority of website users which tend to demonstrate a high intent to purchase. This way, by using Conversion Modelling, you can achieve up to a 20x increase in the number of feedback signals sent to ad platforms, which certainly helps improve the performance of automated bidding alghortims both in Facebook Ads and Google Ads.

What results can you achieve by enhancing your Smart Bidding campaigns with Modelled Conversions signals?

Every business is unique, which means it would not be accurate to promise specific performance uplifts. Having said that, it is not uncommon for SegmentStream clients to report an increase from 50% to 150% ROAS (similar results for CPA-driven businesses as well).

To better understand how Conversion Modelling can improve your marketing performance, we suggest requesting a short trial of SegmentStream’s platform and testing the approach yourself.

What are the prerequisites to use SegmentStream and Conversion Modelling?

As Conversion Modelling is built on Machine Learning, SegmentStream requires at least 3-4 weeks of historical behavioural data, and not less than 300 conversions per month across the entire website, in order to predict the Probability to Convert with very high accuracy.

We also recommend using SegmentStream if your website has at least 50,000 unique website visits per month, and has Google Analytics implemented (Universal, GA360, or GA4).

To summarise,

It is crucial to inform Facebook & Google Ads about the value of clicks they are generating. Without it, ad targeting and optimisation will be limited, and revenue results will be poor.

Using the Conversion Modelling Platform by SegmentStream, you will be able to attribute relative value to each paid click, immediately. By enhancing Facebook Ads & Google Ads campaigns with Modelled Conversions, you will certainly help the ad platform’s algorithms learn faster and target a more relevant audience. Thus, getting more revenue from the same ad budget.

Curious to learn more? Request a trial of SegmentStream to speak with one of our product consultants and see the Conversion Modelling Platform in action.

See SegmentStream in action
Find out the true value of your website traffic sources and boost online ads' performance
Thank you!
We'll get back to you within 24 hours.

You might also be interested in

More articles

Get started with SegmentStream

Learn about Conversion Modelling and why it is a true next-generation solution to outdated marketing attribution and conversion tracking tools.