Today's customer journey is significantly more complex than when single-touch attribution reigned as the exemplary model. Customer journeys are now multi-channel, multi-device, and often non-linear. While more sophisticated attribution models have been developed to address the contemporary customer journey, many marketers and brands still rely on single-touch attribution.
Unfortunately, single-touch attribution cannot provide a complete picture of the customer journey; as a result, seeming insights can be inaccurate, misguiding all optimizations and attempts to grow market share.
Read More
At the highest level, attribution models can be divided into 3 categories:
Single-touch models include first-touch and last-touch, but in each case all conversion credit is attributed to one touchpoint. While limited, these models can be useful in measuring the effectiveness of top-of-funnel and bottom-of-funnel touchpoints, respectively.
There are three main types of rules-based MTA models:
Each model uses predefined rules to determine how to distribute the conversion credit across touchpoints. For example, in a linear model, a customer who interacted with a search ad, an organic visit, and a retargeting email would receive 33% credit for each touchpoint. Linear models help measure the effectiveness of multiple marketing channels in a customer's journey.
Position-based models are useful for measuring the power of both top and bottom funnel marketing elements, while time-decay models are helpful in determining the effectiveness of marketing efforts in driving customers to convert.
Data-driven models give credit driven by the context of all elements of the customer journey, and consider holistically how prospects engage with different touchpoints to become customers. Google Analytics' data-driven model, for instance, may consider data from website visits, ecommerce transactions, YouTube, and Display ads. Ideally, data-driven attribution provides allocating weighting based on the business model, journey context, and the given journey data set. Algorithmic models, which are also data-driven, can use machine learning to automatically allocate attribution weighting based on the complete volume of customer data that trains the model.
The primary objective of marketing attribution is to learn the effectiveness and quantified value of various elements in a marketing mix, such as the channels, touchpoints, messaging, and tactics that drive qualified leads, conversions, and revenue. These components contribute to a company's growth and the successful scaling of operations. However, aligning a brand's marketing mix with its business model, growth strategy, and the actual customer journey is crucial to achieving desired outcomes.
Inaccurate data can lead to misguided insights, suboptimal optimization, and unsuccessful growth strategies. As a result, each business requires an attribution solution that produces precise results, which single-touch attribution models are unable to provide in today's context.
The average number of touchpoints in a typical customer journey vary by industry, as low as 4 for low-cost retail purchases, 12-20 in B2B, and 38 for travel. Ultimately, the number of touchpoints varies greatly by customer segment, complexity of offerings, and even channels. Moreover, a key outcome of an attribution solution should be clarity of a brand’s specific customer journeys. In any case, touchpoints and their influence range far beyond the capabilities of a single-touch model (Think with Google, Salesforce).
Despite these shortcomings, single-touch attribution remains the most accessible model and continues to be employed by numerous marketers. However, its considerable limitations persistently impact the utility of the attribution outcomes.
Many marketers use single-touch attribution because of legacy platform adoption. Whether a CRM, ad platform, or analytics app, many points of entry for attribution are tied to simple, inexpensive, or free platforms. In many cases, early versions of the platforms (Google Ads, Google Analytics, HubSpot) made single-touch the default or easiest attribution option.
Unfortunately, marketers seeing attribution through default or legacy settings are seeing an imperfect and often inaccurate measurement of conversion credit. However, with no accessible alternative reference point, small marketing departments are unable to see their own measurement gaps.
While most modern platforms now offer some version of multi-touch attribution or even data-driven attribution, brands may be slow to adopt new approaches after years of familiarity with “last-click” attribution, for example. In 2021, Google Ads made data-driven attribution the new default. Then, in April 2023, Google announced that most old models will be discontinued in June. Google Analytics and Google Ads will remove first click, position-based, time decay, and linear attribution models. Notably, last click will remain available.
Single-touch attribution models tend to overemphasize the last touchpoint before a conversion, ignoring the other touchpoints that may have contributed to the sale. This can lead to an incorrect credit allocation, limited customer insights, and poor marketing decisions.
All single-touch models share the same flaw: they ignore the contribution of all other touchpoints of customer interaction on a conversion path. However, last-touch models are often attractive due to their simplicity and proximity to the sale. And while the last touchpoint is often the one that triggers the conversion, it is far from the only touchpoint that mattered. Today's customers complete a complex journey, researching, comparing, and evaluating before making a purchase, engaging with multiple touchpoints along the path.
Last-click attribution assumes customers make purchase decisions without any measurable input besides the last touchpoint before the sale. For example, if the last non-direct click is Google Ads, last-click assumes that the ad alone was sufficient to drive the purchase. While small impulse purchases might function this way, other purchases are not predicated only on the last Google Ad before a sale. Products with significant competition, such as espresso makers, laptops, and other electronics, may require weeks of research, numerous sources of information, and a dozen cross-channel interactions before a purchase.
As a result, the business’s understanding of the customer is poor, and drives inefficient if not wholly inaccurate marketing decisions. In the case of highly competitive industries, the content marketing may be the most critical driver of purchase decisions. In many cases, the non-branded content may be the underlying driver of conversions (and traffic), a fact that could be lost with a last-click focus. In such cases, failure to invest in ongoing content development would undercut sales.
Last-click models focus on the bottom funnel only, a small part of the overall awareness to action funnel. While ecommerce commodities might fare well with end-of-journey metrics, large ticket investments or those with lengthy journeys could be devastated.
Overemphasis of last touchpoints is especially problematic when the customer journey is long and complex. For example, in B2B sales, where the decision-making process involves various stakeholders and touchpoints, single-touch can misrepresent the impact of each marketing channel. As a result, subsequent budget allocations may be unaligned with actual channel performance. Similarly, in B2C industries where the purchase cycle is long, such as real estate or automotive, last-touch attribution may ignore the influence of early touchpoints. Brand-building ads and content marketing remain unmeasured when focus is on only the last touchpoints, such as showroom visits or test drives.
More sophisticated MTA models such as linear, time decay, or algorithmic distribute the credit among touchpoints according to their journey position, impact on conversion probability, or other evidence-based factors.
Single-touch attribution models are often confined to a single marketing channel, such as email or display ads, and fail to account for cross-channel interactions. This limitation can lead to overvaluation of certain channels while ignoring the impact of others.
During their journey, many customers interact with multiple channels and touchpoints. For instance, a customer may encounter a social media ad, visit a website, read a blog post, and then sign up for a newsletter. Subsequently, the customer may receive an email and click on a retargeting ad, ultimately leading to a purchase. However, a single-touch attribution model would only credit the retargeted ad, disregarding all other channel touchpoints that contributed to the conversion.
This incomplete picture of the customer journey fails to capture multiple touchpoints and channels, and each touchpoint’s different role. For example, social media ads may raise awareness, a blog post may provide comparative information, a retargeting ad may serve as a reminder, and a promotional email might provide incentive. Each touchpoint contributes to the journey, but single-touch only considers the last (or first) one.
Multi-touch attribution (MTA) models provide a more accurate way to distribute credit based on the role and impact of each touchpoint in the customer journey. Using an algorithmic model that leverages machine learning, contemporary MTA models can consider the idiosyncratic factors of a specific business. By capturing the full cross-channel interactions, MTA models can help businesses gain a better understanding of the customer journey and optimize marketing components more effectively. They can identify which channels and touchpoints are most effective at each stage of the journey, allowing marketers to allocate budget and resources based on specific needs. Moreover, MTA models can identify potential gaps or bottlenecks in the journey, and marketers can optimize strategies and tactics to improve the overall customer experience.
Single-touch attribution models may not be able to track offline interactions, such as in-store visits or phone calls, resulting in an incomplete view of the customer journey.
Even with the rise of digital channels, offline interactions remain crucial in certain high-contact industries such as retail, hospitality, or healthcare. Customers may prefer to visit a store to try a product, call a business for more information or receive direct mail with offers that influence their purchase decision. However, single-touch attribution models would not account for these offline interactions.
For example, a customer may see a digital ad or content, call a business for more information, and finally make an in-store purchase. However, single-touch attribution would only credit the first or last touchpoint, ignoring the contribution of all other touchpoints that led to the conversion.
Multi-touch attribution models can measure both online and offline interactions, assigning credit to all touchpoints. Offline touchpoints, such as calls and visits, can be tracked with offline methods using unique phone numbers, QR codes, or geofencing. ML algorithms can identify patterns and relationships between touchpoints, measuring their impact on conversion probability, whether online or offline.
Single-touch attribution models often allocate all costs to the touchpoint that generated the conversion, rather than to the touchpoints that contributed to the sale. This approach can lead to an inaccurate view of marketing ROI and suboptimal marketing decisions. Accurate attribution is crucial to understanding which channels are driving value, and insufficient attribution models can easily misallocate resources and create imbalances in spends.
Single-touch attribution models tend to overvalue specific channels, neglecting those that contribute to conversions. For example, brands that invest heavily in search ads may over-invest, as they see all conversions through the search lens. When a search ad interaction precedes a conversion, the business may quickly agree with last-click attribution.
Furthermore, a single-touch attribution model does not account for the relative cost of each touchpoint. Costs can vary widely between channels, and businesses must allocate budget accordingly. For example, social media ads may be less expensive than direct mail campaigns, and customer acquisition cost (CAC) should be factored into any impact equation.
Sophisticated MTA models provide greater clarity to the customer journey and identify the efficacy and cost efficiency of each touchpoint. With detailed attribution data, brands can avoid overspending on certain channels while maximizing marketing investment.
Single-touch attribution models fail to recognize touchpoints that played a role in the conversion process but did not generate the final sale. This approach can result in an undervaluation of certain marketing efforts, leading to misleading conclusions and suboptimal decisions. By attributing all conversion credit to a single touchpoint, single-touch attribution models overlook the contribution of other touchpoints with which a customer interacted before converting.
These overlooked touchpoints, known as assists, can play a significant role in driving customers toward conversion. Multi-touch attribution (MTA) models provide greater accuracy in measuring their influence. Both position-based and algorithmic MTA models are more rigorous in capturing the value of all journey elements.
Single-touch models often fail to consider recency and decay in customer journeys, resulting in an undervaluation of touchpoints at both ends of the journey.
Recency refers to the time elapsed since a customer interacted with a touchpoint and can impact its influence on conversion. For example, a customer who saw an ad yesterday is more likely to remember and be influenced by it than a customer who saw the same ad a month ago. Assuming all touchpoints are equally influential or only the closest to conversion is influential can be problematic for complex or long customer journeys. Both first-touch and last-touch models assume equality of touchpoints regardless of recency and do not accurately attribute impact over long purchase cycles.
On the other hand, relying solely on last-touch models can lead to an undervaluation of early touchpoints that played a crucial role in building brand awareness. Complex or lengthy customer journeys must account for the interplay of touchpoints over a long decay window to accurately attribute impact. Learn more from our "Complete Guide to Attribution Models."
Single-touch attribution models fail to consider the long-term value of a customer and the impact of certain touchpoints on future purchases. As a result, businesses may make suboptimal marketing decisions that prioritize short-term gains over long-term value.
Customer lifetime value (CLV) is a critical metric for determining the long-term value of a customer. It considers not only the revenue generated from an initial purchase but also potential revenue from future purchases and the likelihood of customer referrals. CLV looks beyond the immediate conversion and considers the customer's long-term value to the business.
Single-touch attribution models do not account for CLV and assume all customers (and touchpoints) are equally valuable. As a result, they are unable to measure touchpoint influence on the long-term value of customers. MTA models, especially algorithmic models leveraging machine learning, can identify touchpoints with the greatest impact on CLV. By allocating resources toward these touchpoints with long-term impact, businesses can not only improve immediate conversions but also focus on high CLV customers for personalized and targeted content.
With a better understanding of the touchpoints driving long-term value, businesses can make informed marketing decisions that balance long-term customer value with short-term gains.
A key reason to invest in marketing attribution is to scale operations while driving performance. By identifying the highest value market elements, businesses isolate the most valuable investment points to drive customers into the funnel, engage them through purchase decisioning, and pull customers through conversions. However, single-touch attribution models can be difficult to scale to large datasets, which can lead to increased complexity and cost. As a result, single-touch models have limited usefulness for larger businesses or those with complex customer journeys.
In situations involving numerous touchpoints and intricate customer journeys, accurately attributing credit becomes a daunting task. Single-touch attribution models necessitate that businesses manually allocate credit to specific touchpoints, which increases the time and effort expended, as well as the likelihood of errors. Alternatively, brands may resort to relying on advertising platforms to automatically attribute credit to the last ad engaged before conversion, which is equally inefficient. Furthermore, single-touch models generally fail to offer the level of granularity demanded by large organizations with complex customer journeys.
For these enterprises, a comprehensive understanding of the customer journey is essential in order to pinpoint effective touchpoints and channels. Unfortunately, single-touch models fall short in providing the requisite level of detail to be valuable for such businesses. Large organizations with intricate product and service offerings, as well as multifaceted customer journeys, necessitate a sophisticated perspective on their marketing endeavors in order to accurately evaluate effectiveness on a large scale.
For such brands, a machine learning-based MTA model is the most effective option. This model simplifies and automates complex attribution tasks, allowing even large and intricate operations to be measured accurately. Consequently, big businesses can efficiently allocate resources, boost ROI and market share, and avoid manual error.
To extract maximum value from customers, brands must focus on enhancing the entire customer journey and experience. Marketers must strategically invest across the full spectrum of the journey, encompassing all funnel stages. Regardless of how the sequence of touchpoints is assessed—be it broad funnel stages or distinct touchpoints—brands must determine the factors that guide decisions at each stage. Customers can be won or lost at any point, from awareness to advocacy.
Without multi-touch models, marketers cannot fully account for customer experiences throughout the journey. Marketers must identify patterns that drive success and isolate touchpoints where customers are lost. To achieve this, they need a comprehensive attribution model that evaluates the entire scope of customer and brand interactions.
Traditional single-touch attribution models no longer suffice for accurately gauging the influence of marketing initiatives. Even when used effectively, single-touch prioritizes short-term success over long-term growth. Advanced attribution models that encompass the entire customer journey are essential for a holistic view of marketing ROI and facilitating more effective marketing choices for the near and long-term. By embracing a more all-encompassing attribution approach, organizations can secure a competitive edge in their industry and better comprehend their customers.
Recent studies reveal that the journey is more, not less complex, in many sectors. Reductive analytics using templated touchpoints will never match the fluid movement of consumers in a modern funnel.
Ultimately, single-touch attribution inadequately represents the customer journey, leading to errors in measurement, insight, and application. Resource allocation is based on a limited understanding of the customer, jeopardizing the accuracy of all KPIs. Crucial objectives, such as accelerating customer acquisition and scaling for enhanced performance, rely on partial or potentially incorrect data. This results in squandered resources, ineffective insight implementation, and the inability to optimize operations or expand market share.