A Complete Guide to Attribution Models

Cracking the Attribution Code

In today’s world, customer journeys are complex. The proliferation of marketing channels–social, display, email, SEO, print– has led to a fragmentation of the customer journey. Rarely does a sale result from a single ad click anymore. Prospects engage with pieces of content across multiple platforms over time before converting. To properly analyze marketing ROI, businesses need to understand the entire path to purchase.

Attribution modeling provides a solution. As companies seek to optimize their marketing campaigns and allocate budgets efficiently, marketing attribution models have emerged as indispensable tools. These models analyze the customer journey across multiple touchpoints, from initial brand exposure through to final conversion, identifying the touchpoints that ultimately drove the sale. With clear visibility into the conversion path, managers can double down on the highest ROI activities and channels while modifying or eliminating those demonstrating little impact.

Getting attribution right takes work but pays dividends. With clearer views of channel effectiveness, marketers can optimize spending, improve ROI, and drive greater impact on the business. The analytics are within reach, and the hardest part is having the discipline to let data, not intuition, guide decisions.

What Are Attribution Models?

Marketing attribution models are analytical tools that help businesses understand the impact of marketing efforts on conversions. They do this by assigning credit to different touchpoints throughout a customer’s journey. By determining the relative importance of different touchpoints, attribution helps identify the best marketing channels and optimize budget allocation.

How Do Attribution Models Help Marketers

Employing effective attribution models allows brands to see the relative and measurable value of different marketing elements in their customers’ journeys.

  • Identifies the most effective marketing channels and campaigns
  • Enables optimization of marketing investments and budget
  • Quantifies the impact of marketing efforts on sales or conversions
  • Provides insights into the customer journey across touchpoints
  • Determines high vs low value touchpoints in conversion path
  • Uncovers opportunities to improve underperforming marketing areas
  • Supports data-driven and evidence-based marketing decisions
  • Helps understand synergies and cross-effects between marketing activities
  • Improves measurement and accountability of marketing ROI

Ultimately, holistic attribution can drive revenue and reduce waste, leading many to consider the optimization practice as revenue attribution.

Why Different Models?

Each attribution model is designed to answer different questions and provide unique insights. Marketers must identify the right attribution model for their specific business and use case to gain the most actionable insights.

The key is choosing a model aligned with your business’ sales cycle, marketing strategy, and revenue goals. For example, ecommerce sites with short sales cycles can benefit from last-click attribution prioritizing recency. For longer B2B sales with multiple decision makers, first-click or position-based models highlighting early brand impressions are more impactful. And for complex multi-channel journeys, nothing performs better than algorithmic multi-touch models.

Ultimately, marketing attribution empowers businesses to maximize the return on their marketing investment. When aligned with broader objectives, attribution models deliver data-driven clarity into the customer journey, driving smarter channel strategies and higher conversions. In today’s cluttered marketplace, these insights offer a distinct competitive advantage.

Single-Touch and Multi-Touch Attribution

There are two broad types of attribution models:

Single-Touch models, like first- or last-click, credit only one interaction. While easy to understand, these can disproportionally reward certain channels.

Multi-Touch models, such as linear, time decay, and position-based, distribute credit among many touchpoints and provide a more nuanced view of channel interplay.

Single-Touch Attribution Models Explained

Single-source attribution models assign 100% of conversion credit to one marketing touchpoint along the customer journey. First-touch and last-touch are common single-source models, with full credit going to either the first or last interaction prior to conversion. These binary models are easy to implement but risk significant distortion.

When customers see multiple ads and engage with various brand touchpoints before converting, single-source attribution creates false positives and negatives. Full credit is blindly given to one touchpoint, even if many exposures influenced the purchase. This skews insights on channel efficiency and return on investment.

Are Single-Touch Models Dead?

Well, no. For many years, analytics and ad platforms defaulted to single-touch models.  For example, the default for Google Analytics was once last non-direct click. According to recent surveys, over half of the respondent found last-touch “somewhat effective” and nearly half were using a single-touch model.

Single-touch models are enduring because they are simple and built-in to many legacy platforms. Recently Google announced a push for data-driven multi-touch, and announced the removal of many models; nevertheless, Google retained last-click.

While limited, these models can be useful in measuring the effectiveness of top-of-funnel and bottom-of-funnel touchpoints.

First-Touch

First-touch attribution models assign 100% of the credit for a sale or conversion to the first touchpoint in the customer’s journey. This could be an initial ad click, website visit, content download, or other engagement.

First-touch attribution is best suited for identifying early awareness channels and assessing top-of-funnel marketing activities. It works well for companies focused on brand-building or who have short sales cycles.

The key benefit of first-touch attribution is recognizing the outsized influence of early brand interactions. Research shows the first touchpoint shapes customer perceptions and paves the way for eventual conversions, even if the purchase happens much later.

A B2B tech company participates in several large industry trade shows each year. They use a first-touch model to identify booth visitors who go on to request demos. The analysis finds 20% of demo requests came from trade show leads, highlighting the shows’ ability to drive pipeline.

An online retailer of high-end fitness equipment uses display ads to drive awareness and consideration. The first-touch model identifies those who were exposed initially to the display ad. Even though the high-involvement sales cycle is 2-3 months, the model shows the ads planted a seed even if the conversion happened later.

In both these examples, a customer’s first-touch is critical in the sales cycle. In these cases, the model is effective.

Nevertheless, first-touch attribution has limitations. In long, complex sales journeys, the first touch may overstate its impact on the final conversion. This model also overlooks subsequent touchpoints that shape decisions. First-touch risks prioritizing quantity over quality for early interactions.

While useful for understanding initial brand exposure, first-touch attribution provides limited visibility into the full purchase journey. It highlights early awareness channels but disregards the influence of later touchpoints.

Last-Touch

Last-touch attribution models give full credit for a conversion or sale to the final touchpoint in the customer journey. This is typically the last ad clicked, page visited, email opened, etc. before the purchase. Last-touch’s simplicity has helped make it popular, but the limited accuracy is a drawback.

The rationale is that the most recent brand interaction has the greatest influence right before the conversion event. Therefore, last-touch attributes all impact to the final touchpoint.

This model can be useful when conversion paths are short and linear. It works well if the last touchpoint tends to be the trigger for purchase decisions.

An ecommerce company sells products with low price points, around $10-20 per item. Customers typically make quick impulse purchase decisions, with little research ahead of time. The company uses social ads to generate awareness and retargeting ads to engage those who recently visited the site without purchasing. For these impulse buys, the final retargeted ad preceding a sale is the most influential in driving conversions.

While last-touch does identify the final touchpoint correlated with conversions, it provides an incomplete picture. In complex buyer journeys, last-touch risks overlooking other key touchpoints that shape decisions.

Last Non-Direct Click

Last non-direct click attribution gives full credit for a conversion to the final touchpoint that comes from a known marketing channel before a direct visit or purchase.

Non-direct touchpoints include things like email, social media, paid advertising, etc. Direct traffic refers to visits where the source is unknown (like directly typing in a URL).

With this model, if a prospect clicks a paid ad, receives an email, and then directly visits the site before converting, the attribution would go fully to the last ad or email click. The rationale is that known marketing channels guide the customer journey, while the direct visit is just the final step.

A company sends out a series of automated email nurturing campaigns to prospects over time. They implement a last non-direct click model and find 50% of new customers clicked an email before signing up. This shows the nurturing emails play a key role right before conversion.

A B2C e-commerce retailer runs both social ads and newsfeed posts. Using a last non-direct click approach, they determine 20% of purchases had social media as the last non-direct touchpoint. This highlights social’s ability to drive traffic and purchases.

In both examples, last non-direct shows the role of nurturing emails and social media respectively. This provides visibility into which channels are driving traffic and conversions, especially later in the funnel.

Still, like other single-touch models, this model risks overlooking other influential interactions throughout the buyer journey.

Multi-Touch Attribution Models Explained

The limits to single-touch models are many, but the ease of access and understanding keep these models afloat. However, rather than assigning full credit to just one touchpoint like first or last click, multi-touch attribution assigns partial credit to each touchpoint that influences a conversion. This provides a complete, holistic view of the customer experience.

The key goal of multi-touch attribution is not just mapping interactions, but truly understanding the impact of different touchpoints. Analyzing the relative influence and relationships between channels provides data-driven insights on where to optimize.

Multi-touch models identify which marketing activities are most effective at each stage, and how they work together across the entirety of the customer journey. This granular visibility enables smarter optimization of channel mix and campaign coordination.

While more complex than single-touch, multi-touch attribution delivers the clearest picture of how prospects really evaluate brands and make purchase decisions. For today’s fragmented buyer journeys, it is considered the gold standard for mapping journeys and guiding strategic decision-making.

Linear

Linear or even-weight attribution distributes credit evenly across all touchpoints in the customer journey. Every interaction receives an equal share of the credit regardless of placement or perceived impact.

The benefit of linear attribution is its simplicity and inclusiveness. No touchpoints are ignored as potentially insignificant. However, it lacks precision in weighting touchpoints, as not all have an equal effect on conversions.

A software company offers a free trial of their app to generate new subscriptions. It runs ads on Facebook, Google, and industry blogs to drive trial sign-ups. The company also sends periodic emails to trial users to educate them on key features. Roughly 20% of trial users convert to paid subscribers.

In this case, each ad channel contributes incrementally to leading the customer through the journey from awareness to subscription. Then the nurturing emails also play a role in converting trials to subscribers. A linear model would be applicable.

While easy to implement, linear attribution delivers limited actionable insights compared to models that account for real differences in touchpoint influence.

Time Decay

Time decay attribution assigns more credit to touchpoints closer to the final conversion. The influence of interactions is viewed as increasing over time as momentum builds towards purchase.

In this model, the first touchpoint receives the lowest attribution value, while the last touchpoint right before conversion gets the highest value. The logic is that interest and intent accelerate as the customer gets closer to making a purchase decision.

A B2B company sells high-priced software subscriptions to other businesses. The sales cycle is quite long, averaging 6-8 months from prospect to closed deal. Marketing generates leads through webinars, white papers, email nurturing, etc. Sales reps follow-up on leads and guide prospects through the long buying journey. While the early touches are essential, the final sales rep follow-ups are most influential in the ultimate conversion.

In this case, time decay attribution aligns with the customer lifecycle.  However, time decay risks undervaluing early touchpoints that shape brand perceptions and drive initial interest. Not all late stage interactions are more impactful.

Position Based

Position based attribution models assign more credit to specific touchpoints represented by their name. For example:

U-Shaped attribution assigns the highest credit to the first and last touchpoints in the customer journey. The initial and final interactions receive the most weight, while middle touchpoints get moderate credit.

W-Shaped attribution also assigns significant credit to a middle touchpoint. Like U-shaped, the first and last touchpoints receive high weights. But a key middle interaction is given equal importance.

Z-Shaped attribution assigns highest credit to the first and last touchpoints; however, the middle points get partial evenly distributed credit (on a graph, the high points and the middle declining points create a “Z”).

The advantage of position-based is accounting for impact of early brand awareness and final remarketing efforts (U-shaped); or awareness, engagement, and final remarketing (W-shaped); or balances the effect of first-touch, last-touch, and linear models (Z-shaped).

A customer clicks on a display ad, visiting a retailer’s website to browse products. After more research, they subscribe to the retailer’s email list. They open and click on several emails over the next month. Finally, they click on a retargeting ad and make a purchase.

In the example above, a W-shaped position attribution model would credit 30% to the initial ad during awareness, 30% to the middle email subscription, 30% to the retargeted ad. The remaining 10% is distributed to the multiple mid-journey email engagements.

Data-Driven or Algorithmic

Data-driven attribution leverages AI and advanced analytics to determine the optimal credit weighting for touchpoints based on data. These data-driven models analyze historical customer journey patterns to identify the true influence of different interactions.

Rules-Based vs Data-Driven Attribution

While the industry usually divides all attribution models into single-touch or multi-touch, others draw a distinction between rules-based and data-driven attribution. Rules-based attribution models assign credit to touchpoints according to predetermined rules rather than algorithmically derived weights.

Common rules-based models include first-click attribution, which assigns 100% credit to the first touchpoint. More advanced rules-based models, like linear or time-decay attribution, aim to better distribute credit across channels but still rely on simplistic rules rather than data-driven optimization.

Data-driven attribution models are the outlier of attribution models. Powered by advanced statistical techniques like Markov Chain and Shapley Value modeling, these algorithms analyze huge volumes of customer data to uncover the nuanced influence of different touchpoints on conversion.

Rather than using predefined heuristic rules like other models, machine learning attribution derives channel value solely from performance data. The algorithms detect subtle correlations between touchpoints and conversions across large volumes of data.

The upside is significant: strategic insight on channel efficiencies, informed budget allocation, and optimization of the consumer journey.

A customer sees a brand’s social media ads several times over the course of a month, building awareness. They visit the brand’s website and download a couple pieces of content. A few weeks later they fill out a contact form to request more info. After a sales call, they sign up for a free trial. The customer extends their trial once before finally purchasing.

In this journey, there are multiple touchpoints across digital ads, content, and sales interactions over an extended period. The influence of each touchpoint is complex and nuanced. Simplistic attribution models would struggle to properly weigh each interaction. But data-driven attribution would analyze large volumes of historical journey data to detect correlations between specific sequences of touchpoints and conversions.

The machine learning algorithms would determine, for example, that early ad impressions have a subtle but measurable impact on eventual conversion even though they are temporally distant. Or that users who download certain content assets convert at higher rates. Note, that even Google has placed their faith in data-driven attribution for both GA4’s DDA as well as Google Ads modeling.

By processing huge amounts of granular data, data-driven attribution can account for intricate relationships and precisely optimize spending and activities for maximum impact throughout the funnel. No preset rules can match its rigor.

Choosing an Attribution Model

There is no one-size-fits-all approach to attribution modeling. The most appropriate model depends on marketing objectives, sales cycles, channels used, and data sophistication.

For short, simple funnels, single-touch models may suffice to identify key interactions. But for multi-channel journeys, multi-touch attribution provides a more nuanced perspective on cumulative touchpoint influence.

The key is aligning the model design with how your business engages customers across channels and campaigns. Ideally, a business must define success metrics, analyze historical interactions, and test different model types. Often a hybrid approach works best, blending single-touch simplicity with multi-touch comprehensiveness. Set stipulations and custom weightings tailored to your strategy and data.

Finally, a multi-model platform agnostic to specific models allows brands to view both an unfiltered customer journey as well as performance under different scenarios. For example, Arcalea’s multi-touch attribution platform, Galileo, allows the business to compare individual and aggregate customer journeys using raw data or various attribution models. Moreover, using the Scenario Planner, the platform can predict return for reallocated ad spend.

As attribution modeling matures, leveraging AI and advanced analytics will enable more intuitive mapping of complex journeys for more brands. Still each brand occupies a different space in the data maturity model. Start where you are. Investigate software vendors and attribution partners. Define your attribution roadmap. And begin to grow your knowledge of the customer journey and the impact of your marketing mix.