a chart to help show how conversion tracking is helping

Conversion Tracking: More Than Most People Would Ever Want To Know

Conversion Tracking: It Means Someone Somewhere Did Something, & We Like It

Conversion tracking is designed to measure the results of online campaigns by tracking specific user actions, aka “conversions.” At its core, conversion tracking relies on data from a website or app being sent to analytics and ad platforms. This data is attributed back to specific ads, targeting segments and other sources of the visit, allowing marketers to have smarter reporting and optimized campaigns.

There’s not a single approach to conversion tracking. Not only is the implementation usually quite different from site to site, but the way data is interpreted and how value is calculated can vary. The way attribution is calculated to begin with can also change based on the analysis goals.

While conversion tracking is intended to be an objective measure of a campaign’s results, the way it is interpreted can require nuance. Understanding the full picture of conversion tracking is an essential skill for savvy marketers.

Here’s How Conversion Tracking Might Be Explained To Different Peeps:

Please forgive any weirdness in these examples – doing six of them took a toll.

Imagine you have a lemonade stand and people either drive up or walk up. Every time someone buys a lemonade, you put a star on a board to count how many sales you made.

To help you know more about your sales, you upgrade to colored stars. You put a blue star on the board when you sell to a truck, a red star when you sell to a car, and a green star when you sell to someone who walks up.

Now you are counting not only your sales (we would call them conversions, or goals) but also the blue/red/green stars show you where your sales came from based on how they arrived at your lemonade stand. In the future you might use this data to see if changing your sign or location increases your lemonade sales overall, or if it increases for Cars vs Trucks vs walkups.

Imagine you’re planning a party, and you want as many people as possible to come. You try different ways to invite people: you post on Instagram, send out a group text, and make a video to share on Snapchat.

Conversion Tracking is like knowing which invite method worked best. Knowing what method actually gets people to show up lets you be more effective in how you reach out to different “channels” like posting on social platforms or texting.

We can analyze more data too, like how much time each method takes compared to its result rate. Perhaps making a video on Snapchat has the best success percent, but it also takes you 10x longer than sending a group text. We analyze conversions not just in terms of conversion rate but also in terms of cost versus value.

Imagine you’re running search ads for your business selling Widgets, and you’re investing time and money into it. Conversion tracking is the tool that helps you see whether those efforts are actually paying off, resulting in sales and actually making you a profit.

You set up to tracking specific actions people take on your website after interacting with an ad, like making a purchase (the main goal), signing up for a newsletter (which may lead to a later sale), or contacting you for more information (maybe they just have a question before they buy, or maybe they want to place a bulk order).

You pass the conversion event as well as the value back to the search ad platform, which then provides you and the platform critical data to understand cost vs value. In a week, you spent $100, sold 50 Widgets for $10 each and you made $500. You stuck $100 into your magic marketing box, and turned it into $500 (that’s $400 profit). Great! This gives you the confidence to know you’re making money and should continue your efforts.

To continue the last example, you also know that 10% of newsletter signups result in a Widget sale, so you can actually assign value there too. A 10% likelihood of a $10 sale means a newsletter signup is worth $1 on average. This gives us two different conversions we can focus on, both with an assumed value.

In the original campaign, we turned $100 into $500 via direct sales (we’d call this a 5 ROAS). In a separate test campaign that spent $100 and generated only 20 sales with a total value of $200. We’re upside down, oh no!

But we also see the campaign generated 250 newsletter signups. Based on our newsletter to purchase conversion rate, we may not have recouped the sales immediately but we have now actually have a campaign that generated $250 (email) + $200 (the direct purchases). So in effect, that test campaign generated $450. Some of this is assumed future value, so we want to continue to watch the newsletter conversion rate to purchase. $100 -> $450 is a 4.5 ROAS, so it’s somewhat lower but this is additional revenue through a different avenue.

Important: different campaigns can satisfy different business goals and have different levels of efficiency.

Conversion tracking quantifies efficacy of digital marketing by capturing user interactions that correlate with business objectives, such as transactions, lead submissions, or other key actions. Conversions may be a primary goal macro. Alternatively, it might be a micro behavior that indicates higher funnel interest, which may be a future converting user. Campaigns may specifically optimize towards generating and capitalizing on individuals who send each of these conversion signals. Data may come from tracking pixels, script-based event tags, uploaded manual events matched with user ID tools, or even server-based conversion tracking. Detailed attribution associates these actions to specific campaign elements, from ad creative to placement.

User identifiers monitor behavior across touchpoints and ad exposures, enabling multi-touch attribution models to evaluate not just the final touchpoint but also intermediate interactions that initiated or contributed to a conversion. Full conversion path understanding is critical to ensure you understand initiating sources that may show as assist conversions or conversion touches, but may not have earned attribution based on the particular conversion model.

Understanding of conversion models and how click/view attribution behaves together is important in order to see the various ways we could understand value contribution from different sources. Without this deeper understanding of attribution, optimizing away a “low performer” might inadvertently decrease your upper funnel incremental conversions, which may have a longer consideration cycle.

Conversion tracking, in its most technical form, leverages user behavior data to construct detailed, actionable pathways from initial engagement to final conversion. We consider data from individual touchpoints such as what creative & targeting has contributed a touch event, to entry channels on a website, to what onsite elements they click on. We try to understand not only the value of all involved sources/segments, but also how each of these individually or as part of a conversion path influence the propensity for a user to perform additional valuable actions.

Like our prior example, we provide conversion value across various touchpoints to ad platforms to monitor results and provide cost to value insight. But! Now with our very deep understanding of conversion data and how platforms leverage machine-learning algorithms based on this data, we can shape it to our advantage. Platforms need sample size and we can fill in the gaps.

Given our knowledge about micro conversion events, conversion touches and other high-funnel behavior’s propensity to convert (and the effective projected value weighted by the micro conversion’s likelihood to produce future value), we can feed ad platforms our model of goal behavior that represents the projected value of future conversion events.

In essence, not only would we tell ad platforms about a purchase, but we inform them about all non-purchasers who might purchase – based on how they interact and what we can project their purchase proclivities are based on prior performance. A savvy marketer develops and monitors this model so ad platforms are able to learn from real-time behaviors across the conversion path.

Tagging for Conversions

What It Is: A conversion tracking pixel or script is a small piece of code added to your website that allows ad platforms to track user actions and measure specific events. This might be hard-coded on the website (not ideal) or it could be implemented via a tagging tool, such as Google Tag Manager (very ideal). This kind of tool defines what conditions will trigger a tag to fire, and then fires corresponding tracking tags to various ad platforms that need to be informed of the conversion event.

How It Works: When a user engages with an ad and lands on your website, the pixel or script collects data on their actions such as page visits, purchases and other events. It often uses cookies or unique identifiers to recognize users across sessions and devices. Tracking tags not only collect the Macro conversion information (the main data), they also track Micro conversions (minor interactions generally leading up to conversion). Other information that can be tracked includes metadata, such as product IDs, prices, and user session details, helping platforms understand which actions are valuable conversions.

Information Conveyed: Tracking tags communicate based on IDs or Hashed PII, meaning as advertisers we’re not directly receiving the email, phone number or otherwise private information of someone who is being tracked. The gist of what we’re seeing is: User123 engaged with an ad, and User123 also was seen on the website buying a Widget. Therefore, the campaigns and targeting segments that had exposure to User123 are given conversion credit, without needing to know any specific information about that user.

What Ad Platforms Do With Conversion Information

Attribution and Optimization: Once ad platforms receive conversion data, conversion attribution is calculated, assigning these conversions to specific ads, keywords, targeting, ad groups, audience segments, and other advertising criteria, enabling marketers to see what’s driving results.

Performance Optimization: Ad platforms use conversion data to refine and optimize ad delivery. For example, if Facebook’s algorithm sees a specific ad leading to conversions, it might prioritize showing this ad to similar audience segments. Google Ads might adjust its Automated Bidding system’s bids for high-converting keywords or prioritize certain responsive ad elements for optimal performance. Not only do (human) marketers need conversion information to make informed decisions, modern advertising technology is highly adapted to understanding what constitutes conversion value and assists in optimizing. For this reason it’s imperative that conversion tracking information is correct and complete.

Reporting and Insights: Ad platforms provide analytics on conversion performance, enabling marketers to make data-informed decisions about ad spend, creative, and audience targeting. While the ad platforms and marketers may take direct action based on various attribution models and nuanced approaches to conversion data, reporting platforms will provide summary conversion information – and often generalize reported data based on last interaction to standardize how reporting is assigned.

Advanced Concepts: Attribution & Models

Attribution: Attribution is the process of identifying which channels and touchpoints contribute to a conversion. It answers the question, “Which ad or channel gets credit for this conversion?” and helps marketers understand the customer journey across multiple touchpoints. A conversion may be fully or partially attributed. Ex: a purchase may report as one single conversion, but this conversion may also be split into fractions of a purchase based on the campaigns or strategies that participated. How it’s split is defined by a model.

Attribution Models: These are rules or frameworks for distributing credit across different touchpoints. Common attribution models include:

  • Last-Click or Last Impression: Gives full credit to the last touchpoint before conversion. This is a very common default way for reporting of conversions. Essentially, what is the last ad someone engaged prior to taking goal action?
  • First-Click or First Impression: Assigns credit to the first interaction. This is often used to understand what the point of initiation was for a conversion – what was the starting point of the conversion path?
  • Linear: Distributes credit evenly across all touchpoints. This might do a good job at understanding the frequency of other touchpoints as it will cause uneven conversion credit when a particular touchpoint occurs more often for a user, but linear weighting hides the value of key moments in a conversion path.
  • Time Decay: Gives more credit to touchpoints closer to the conversion. This distributes conversion credit favoring recency, which may be valuable to some advertisers but this method hides the initiating component of the conversion path by deliberately devaluing it.
  • Data-Driven: Uses machine learning to assign credit based on each touchpoint’s impact. Data-driven models are often used to try and reflect real, varied user behavior. However, the way this model functions may not be fully disclosed by ad platforms – additionally the weighting applied may vary from converting person to person.

The Need For Testing

Don’t Get Complacent, Stuff Breaks. Websites change, platform requirements change, tech gets updated or deprecated (or there might actually be conversion tracking gremlins), or even platforms like Google will tell us the tracking world is ending unless you Hash PII and implement Enhanced Conversion Tracking before backtracking – anyways…

From time to time it’s valuable to do your due diligence. Run through your conversion steps with your tagging tools in debug mode or monitor the raw network traffic as you fire sample conversion events. Do real test transactions and then refund them. Test your tracking setup and look for any gaps. Validate your data.

Exercise common sense too – don’t believe your data until you’ve confirmed it. If something looks too good to be true, it’s more likely that something is broken. A campaign shows a 100% conversion rate? That would mean every click is converting – not they are not. Something is very broken, like a “marketer” defined the landing page view as the conversion and they deserve shame. Fix it.

Give your ad platforms the quality data they need.

Additional Advanced Topics & Further Reading

Deeper Analytics: Tools like Google Analytics 4 and Facebook Ads now use predictive analytics to identify high-converting user segments and potential lifetime value (LTV), enabling more refined audience targeting. These may also provide information about the presumed Audience or Cohort of converting traffic. This may open up new avenues to reaching individuals similar to your converting traffic.

Multi-Touch Attribution & Cross-Channel Tracking: With users frequently switching between devices and channels, cross-device tracking solutions attempt to unify their journey, enhancing attribution accuracy. This approach considers all devices involved in various touchpoints along the user’s journey and assigns weighted credit to each interaction based on its contribution to the conversion.

Offline Conversion Tracking: For businesses that interact with customers both online and offline, platforms can offer tools to import offline conversions (e.g., in-store purchases) and link them to online ad interactions. There are various ways this is done, and it’s commonly addressed in summary as platforms don’t like to be very specific about how they attach in-person behavior to digital profiles, but in essence your email or phone (or visit) to a retailer can be attached based on matching these signals to your device, and understanding if your device has received ads. It’s similar to typical conversion tracking tags and pixels, but has additional matching steps (and data loss potential).

Server-Based Tracking: Similar to offline conversion tracking, server-based conversion tracking works in the same regard, but is generally done on websites and apps at the time of a conversion (or in batches later). This does not rely on conversion tracking pixels or scripts and is instead direct communication between servers, alleviating tracking issues caused by ad blockers (which can block pixels, but cannot block server communication). Pixel/Script-based conversion tracking is typically deployed in parallel with server-based, allowing marketers to understand match rate and analyze consistency between traditional measurement and server based. This can allow modeling data gaps and establishes a more complete picture. Server-based tracking can be a more complicated way to establish conversion tracking, but has sizable benefits.