Resource Type: Article

Multi-Touch Attribution, or MTA, seeks to better understand what truly drove or “caused” an action to happen. How does multi touch attribution work? Usually, these actions are online and commercial in nature. The most typical use case is understanding what drove an order or a transaction. These “part-worths” can then be summed up across many orders or transactions to form a complete picture of how different marketing channels contributed to the sales accrued in a given period.

For example, say a user purchased a subscription to a monthly ride-sharing service. It is easy to know that the user entered from a paid search ad; each ad has a unique referring URL that can be tracked. However, giving all credit to the paid search ad is most likely unfair. The user might have seen a social media campaign days before, or an ad on a baseball game weeks before. Those touches should also receive credit. The difference between a last-touch and multi-touch attribution is usually substantial. Lower-funnel channels usually look less attractive in a multi-touch approach (see Figure 1 below).

Figure 1: A last-touch versus a multi-touch approach can yield very different results. The “MTA effect” shows less attractive results for paid search and affiliate, and better results for DM, online video, and social in this example.

To accurately apportion credit to all contributing channels, it is necessary to estimate the effect that each touch had on the customer who ultimately ordered. This requires, at the most basic level, knowing that a previous interaction occurred. This is the number one problem with multi-touch attribution: gathering the “touches”.

Historical Context of Tracking

In the first decade of the 2000s, tracking users across digital properties was easier. Customers’ browsers held records of where they had been and what they had done. These cookies functioned as digital breadcrumbs, alerting an advertiser that a consumer had seen an ad weeks before. It was possible for advertisers to piece together customer journeys deterministically—that is, knowing concretely when and what someone saw before they made an order. With these data censuses, early MTAs could estimate each touches’ effects simply—typically applying even credit to each touch, with linear decays of impact over time—or using more complex methods.

Privacy Measures Impacting MTA

However, over the past several years, it has become more difficult to build truly deterministic MTAs. Privacy measures—whether cookie deprecation, walling off app data interchanges, or obfuscation of PII  in third-party data sources—have made it harder and harder to say that a pageview or click was the result of an individual customer with any degree of confidence. The tradeoff between privacy and customization will likely be a forever war (see Figure 2)—suggesting an owned, decoupled approach to attribution that doesn’t rely on fads or black boxes.

Figure 2: The push and pull between privacy and personalization will likely be a forever war.

First-Party Data: The Advertiser’s Advantage

Some data sources are still totally knowable. Anything directly controlled by the advertiser—sometimes called first-party data—can be tracked in principle. Examples of these first-party interactions include email, direct mail, landing pages (or any web pages on the company website), owned or cooperative advertising networks, and call centers—whether customer support or acquisition-focused. In these cases, previous touches and interactions with an individual can be reconstructed as an identity graph, tied together with some unique ID.

Dealing with Unknown Identities

In most cases, however, interactions with customers are “underwater”—unknown to everyone but the individual and the publisher that he interacted with. In these cases, all hope is not lost. There are two basic approaches to dealing with resolving unknown identities.

Identity Resolution Technology

The first is using an identity resolution technology provider. These data brokers provide crosswalks between different walled gardens, by either inking proprietary sharing agreements or by using algorithms to match up individuals by using various pieces of their identity to build “fingerprints.” The fidelity of matching can vary from very low to approaching 50%—which is quite good. The trick is, most of these providers do not want to give up their raw data. Instead, they prefer to keep identities in their domain, for obvious reasons. In fact, most identity resolution providers’ first business is activation, not measurement. They provide value by essentially extending retargeting—the continued advertisement to one user over time—to broader parts of the digital domain.

Even so, it’s worth following up with a few vendors to see what they are willing to provide. In some cases, APIs may be provided that return a proprietary ID (or not) to a given third-party interaction. At scale—over millions of interactions—this approach can help build a deterministic identity graph.

Probabilistic Assignment of Interactions

Another approach is probabilistic assignment of interactions. This approach—while less precise than deterministic assignment of interactions—is also not vulnerable to future changes to platform privacy policies. In this framework, each identity (user) is assigned a probability of having seen an advertisement based on time and location.

For example, say we know that in the Memphis DMA on May 1^st, there were 1,045 paid social ads served (these types of data are usually available from platforms, or via the digital agency executing the marketing.) The only other piece of data we would need would be the targeting criteria for the campaign (say, 18-34 year olds), and the number of those people in the DMA. From the American Community Survey (ACS), we know that the 2022 1-year estimates for 18-34-year-olds is 367,709 in the Memphis MSA (Metropolitan Statistical Area)—not an exact match for DMA but close enough. In that case, we can say that on May 1^st, an 18-34-year-old individual in that area would have a 1,045 / 367,709 (0.28%) chance of having seen that ad.

This approach can work for virtually any media channel, from targeted to broad reach. It obviously is less accurate and contains significant potential error, but in aggregate, it is a powerful tool to create all encompassing media attribution.

Building a Longitudinal Human Record (LHR)

In reality, a mixture of discrete (identity resolved) and aggregated (estimated) impressions will be used in a comprehensive MTA (see Figure 3). Whatever the combination, the aim is to end up with a longitudinal human record (LHR)—the data structure also used for CDPs (customer data platforms.) This data structure in “tall and skinny”—in other words, it will have many rows and few columns. A typical LHR used for multi-touch attribution will have hundreds of millions or billions of rows, and something like 10-20 columns.

Figure 3: A mixture of record-level (deterministic) and aggregated (probabilistic) data sources to build the longitudinal human record (LHR)

The most important columns in the LHR are the date-time of the interaction in question and the ID of the “individual”. The records should be sorted first by ID, and then by date-time. In other words, all interactions associated with an ID x will be grouped together, and then sorted by date-time descending, with the most recent interaction at the top. In the case of an order—the thing you are trying to attribute in MTA—this will generally be at the top of a lot of previous interactions.

Apportioning Value to Interactions

Once the data are constructed, a method is required to apportion value to each of the previous interactions. The simplest method is equal apportionment—and it’s not necessarily a bad place to start. Say an ID associated with an order has 40 logged interactions in the prior 90 days. In this case, each of the 40 interactions would receive 1/40^th of the credit for the order.

This simple method can be made more accurate by discounting “probabilistic” interactions by their probabilities. For example, in the case of the 0.28% chance someone in Memphis saw the social ad on May 1^st above, that record would be a 0.0028 in both the numerator and the denominator of the attribution equation. It follows that deterministically known records—for example, an email that was clicked—will get a 1.00, which will swamp the probabilistically estimated interaction (or impression.)

Advanced Methods of Parsing Credit

More advanced methods of parsing credit are certainly possible. There are two primary methods: logistic regression at scale and Markov Chain estimation.

Logistic Regression

Logistic regression at scale attempts to estimate each prior interaction’s relative impact on an event (usually a sale) by transforming each side of the regression equation with a logarithm. The advantage of this approach over more traditional linear regression is that because the input (independent) and output (dependent) terms are logged, any combination of independent variables is bounded between 0 and 1 (or 0% and 100%). This is also called modeling the log of the odds of an event (the transaction). It is out of the scope of this paper to get into the specifics of how logistic regression (or logit) models works, but suffice to say that the output will be a term for each potential input expressed as a log of the odds, which can be translated to probability with the equation:

Where t is the linear combination:

The trouble with using logistic regression for MTA is that it requires ones (successful sales) and zeros (unsuccessful sales—or potential customers who never became customers.) Unsuccessful sales can be hard to come by. When an entire customer journey is known, for example in the case of a health insurance carrier attempting to get members to take a discrete, known action, this is a very feasible approach. However, when prospective customers don’t have a real defined endpoint, more creative approaches to defining zeroes are needed.

Markov Chain Estimation

The second more advanced approach to parsing value in Markov Chain estimation. This flips logistic regression on its head, and focuses on the upstream tactic—say, clicking an email—and then estimates each next downstream outcome possible from that point. A good visualization of this approach is the old “The Price is Right” game “Plinko.” In this game, the player places a disc at the top of a board with a set of pegs. Each peg represents a random coin flip—so the path that the disc takes is a series of .50 probability decisions resulting in an outcome. However, in Markov Chains, each probability is not 50%–and the disc can end up in more than two next states in the next step.

Markov Chains have the fundamental property of “memorylessness”—their behavior at a given state is only dependent on where they are, not where they have been. This simplifies MTA analysis, but by implication, does not take into account the “saturation” effect of marketing. In other words, we only care that you received a piece of mail at time T, but not how many pieces you received prior to time T. The task of the analyst is to build a matrix of the probabilities of the interaction a customer will have next at a given interaction point, sometimes called a transition probability matrix. For example, this matrix might look like this for an advertiser deploying mainly digital tactics:

		Next State
		Initial	Email	Social	Affiliate	OTT	Sale	No Sale
Current State	Initial	0	0.40	0.20	0.30	0.10	0	0
	Email	0	0.10	0.05	0.05	0	0.01	0.78
	Social	0	0.02	0.01	0.02	0.02	0.02	0.91
	Affiliate	0	0.04	0.01	0.03	0.01	0.03	0.88
	OTT	0	0.04	0.01	0.01	0.02	0.02	0.90
	Sale	0	0	0	0	0	1	0
	No Sale	0	0	0	0	0	0	1

The above matrix must meet two conditions:

• P_ij is the probability of moving from state i (row, current state) to state j (column, future state); it must also be bounded between 0 and 1 (inclusive)
• The rows (i) have to sum to 1; sale moving to sale is called “absorption” (yes, this is a bit confusing, you can think of it as the journey ending)

From this matrix, it’s possible to calculate the absorption probabilities of each cell—in other words, the probability that at that point, an individual will eventually turn into a sale. To do this, split thetransition matrix into two sub-matrices, Q, the matrix of transitions between non-absorbing states (touchpoints), and R, thematrix of transitions from non-absorbing states to absorbing states (conversion or non-conversion).

Matrix Q:

	Initial	Email	Social	Affiliate	OTT
Initial	0	0.40	0.20	0.30	0.10
Email	0	0.10	0.05	0.05	0
Social	0	0.02	0.01	0.02	0.02
Affiliate	0	0.04	0.01	0.03	0.01
OTT	0	0.04	0.01	0.01	0.02

Matrix R:

	Sale	No Sale
Initial	0	0
Email	0.01	0.78
Social	0.02	0.91
Affiliate	0.03	0.88
OTT	0.02	0.90

With these two matrices, absorption probabilities are found by calculating the fundamental matrix, F = 1 – Q)^-1 where I is the identity matrix and (1 – Q)^-1 is the inverse of the matrix I-Q. Finally, multiplying F by R will produce the ultimate probability (absorption) of a sale happening (or not) at any point in the matrix. In the case of the given data, F for “sale” is:

		Next State
		Email	Social	Affiliate	OTT
Initial State	Email	0.011151	0.000569	0.000587	0.000176
	Social	0.000488	0.020235	0.000447	0.000418
	Affiliate	0.001401	0.000387	0.031011	0.000324
	OTT	0.000925	0.000256	0.000263	0.020416

These estimates can then be used to allocate value between touches in a chain. For example, say a user is observed going from email to email to social to OTT, and then finally closing. In this case, the values pulled from the above table would be 0.011151; 0.000569; and 0.000418, with the final step—OTT to absorption—pulled from our first matrix as 0.02. To apportion value, these numbers are summed and divided by the total number of steps in the pathway:

(0.011151 + 0.000569 + 0.000418 + 0.02)  =  0.03214

To get to a final contribution vector for each unique journey:

Touch 1 (Email)	Touch 2 (Email)	Touch 3 (Social)	Touch 4 (OTT)
35%	2%	1%	62%

Of course, this has to be scaled over thousands and millions of orders—and it’s a lot more complicated than the simple example shown above.

The final question is time. In the example above, say the first interaction (email) happened 90 days ago, but the last interaction before close (Social) was two days ago. Is it fair to give the Email touch 35% credit and the Social touch 1%? Almost certainly not.

There are many techniques for decaying a touch’s impact on ultimate conversion, but they all have the same basic concept: an interaction’s impact fades over time. In MMM (media mix models) we use the term adstock—the rate at which a touch’s “power” decays after it is shown. The concept is similar in MTAs, but is reversed, because we are focused on the sales, not the stimulus. In other words, we look back from the order, not ahead from the stimulus.

The simplest approach for MTA is straight-line decay with a look-back window. This is just as it sounds: A touch before the start of the look-back window gets 0 credit; one mid-way through the window gets ½ of the full credit; and a touch on the day of the order gets full (1) credit. More advanced logistic decay approaches are certainly possible, but yield limited additional benefits and add significant complexity.

Bringing It All Together: How Does Multi Touch Attribution Work?

Putting all of these elements together—the longitudinal human record (LHR), probabilistic inference of non-deterministic touches, estimation of contribution via either logistic regression or Markov Chain analysis, and decay—will give a marketing analytics team the basic elements of multi-touch attribution.

One final, important note: black box MTAs provided by vendors claiming a secret sauce for identity resolution are compelling because they seem simple, but buyer beware. Proprietary identity graphs are only as good as the federations the company belongs to and the underlying cross-walk code, and there is really no way to validate them other than by seeing if they “make sense.” For this reason, they tend to struggle over the long run as accuracy—and hence utility—are questioned by financial decisionmakers. A purpose-built, open-code, white box approach to MTA, which can also be used to power downstream applications like dashboards and econometric MMM panels—should be the preferred approach for marketing analytics teams.

In the ever-evolving landscape of marketing attribution, businesses are engaged in an arm race to better measure and maximize ROI and growth. Beyond simple last-touch attribution, the two prominent methodologies often compared are Media Mix Modeling (MMM) and Multi-Touch Attribution (MTA). This article delves into the nuances of MMM vs MTA, exploring their differences, advantages, and how they can be effectively utilized to drive marketing success. The good news is that this isn’t necessarily an either / or decision; MTA can be used upstream and in concert with MMM to combine quick-read and long-run views of marketing’s effectiveness. 

Before diving into MMM vs MTA, it’s important to note that the bread and butter of marketing reporting—last touch attribution—is still a powerful and simple tool, and should not be cast aside in favor of more advanced techniques. Last touch attribution credits the final interaction before conversion, generally using a direct database linkage via some kind of unique ID—a telephone number, URL string, or webpage tag. Last touch attribution will always be simpler and faster than multi-channel techniques. MMM and MTA outputs should always be looked at side-by-side with last touch reporting. It is often these comparisons that yield the most interesting insights. 

Understanding Media Mix Modeling (MMM) 

Media Mix Modeling (MMM) is a statistical technique used to estimate the impact of various marketing channels on sales performance. It aggregates historical data in a time series, usually across a geographic or other cross-sectional key, to measure the effectiveness of marketing efforts and to allocate budgets more efficiently. MMM can interpret any kind of stimulus, whether paid or earned, upper- or lower-funnel, or offline vs. online. It can also be used to understand the impacts of non-promotional factors—including price changes, competitive actions, product launches, and distribution channel strategy. 

Figure 1: The basic idea behind MMM; there is an efficient frontier for marketing achieved by optimally mixing channels. This mix is different at different spend levels, but generally the macro curve exhibits diminishing returns to scale (its slope decreases as a company spends more). 

Key Benefits of MMM 

• Comprehensive View: MMMs provide a broad and complete overview of how different marketing channels interact and contribute to overall sales. This comprehensiveness is beneficial to understanding the combined effects of multiple marketing efforts and avoids over-crediting.  
  
• Long-Term Abilities: One of the strengths of MMM is its ability to account for the longer-term impacts of marketing—whether over weeks and months, or years, in the case of marketing’s impact on brand equity. This is particularly helpful when trying to gain an honest accounting of the effectiveness of upper-funnel channels like TV and print, where effects are usually not immediate. This long-term focus also makes MMM less capable of reading “what just happened”—although techniques like rolling analysis windows can help with trending. 
  
• Requires “Smaller” Data: MMM data frames are only a few megabytes, usually a few thousand rows and a few hundred columns. This generally makes it possible to store the data on a traditional device—a hard drive or simple cloud-based storage. MMMs can utilize either spend or impressions as its “x” or independent variables, and can use either sales or revenue as an ultimate dependent variable. Even so, MMMs need extensive historical data, usually a minimum of two years, making them more suitable for established brands with significant data accumulation. Brands without much historical data can still be boot-strapped with Bayesian priors. 

Understanding Multi-Touch Attribution (MTA) 

Multi-Touch Attribution (MTA) is a more granular approach that focuses on assignable channels—usually digital, but also including direct mail and interactions with known customers. It attributes conversions to the multiple touchpoints that a consumer interacts with throughout their journey. This method provides insights into the effectiveness of each touchpoint in the conversion path.  

MTA was very popular in the early days of digital marketing, before privacy concerns and platform data hoarding made it harder to resolve identities across channels. In recent years, it has come under fire as first generation deterministic ID resolution approaches failed, but advances in data clean rooms and probabilistic exposure inference are making “second generation” MTA models a very attractive option for inference. 

Key Benefits of MTA 

• Close to Real-Time Insights: MTA models are capable of using real-time data, allowing marketers to make quick adjustments to their strategies. This is particularly advantageous in fast-paced digital environments. The devil is in the details, however—real-time results are only possible with very robust data pipelines and fast compute environments. 
  
• Potentially Unlimited Granularity: Because MTA models are built at the log level, there is potential unlimited detail available about each individual touchpoint, helping marketers understand the specific role each interaction plays in driving conversions. Keep in mind that this detail is dependent upon robust lookup tables and cross-walks, as well as thought through marketing taxonomies. 
  
• Consumer Journey Mapping: A side benefit of building the human longitudinal record required for MTA analysis is a 360-degree view of the journey. Exploratory data analysis (EDA) of this data artifact using big data tools can identify influential touchpoints, find breakdowns in e-commerce pipelines, and discover high-value audience segments. Even so, MTA data frames store much more data, at the record level. Log files, sometimes billions or tens of billions of rows, must be processed, demanding a big data compute environment. 

Combining MMM vs MTA for a Holistic Approach

MMM vs MTA is how these methodologies are often perceived, but they can and should be used together. Integrating the macro-level insights from MMM with the micro-level details from MTA can provide a comprehensive understanding of marketing effectiveness. This integrated approach allows businesses to leverage the strengths of both models. 

The record-level data required for MTA analysis can be used upstream of the MMM econometric panel structure, directly feeding it. In this way, the same “single source of truth” can be used for both analyses. Data that is not used in MTA—for example, survey data—can be joined after the raw data has been grouped and aggregated. Our recent white paper on the Go-to-Market Data Lake architecture details this approach. There are three main steps: 

1. Creation of Longitudinal Human Record (LHR): Tying customers’ journeys together in longitudinal chains can help locate points of friction, profiles audiences, and conduct multi-touch attribution (MTA). 

2. Creation of Econometric Panel: The LHR then serves as the base query to create an econometric panel for MMM. This panel is a summation of stimulus (x-variables) and response (y-variables) by day or week, across one or more cross-sectional dimensions. 

3. Data Aggregation and Supplementation: The panel is then supplemented with aggregated data, such as linear television or unresolved digital marketing data, to fill in gaps and ensure a complete dataset. 

Figure 2: Start with record-level data to build the LHR, and feed the ultimate econometric panel to enable MMM. 

Use Cases for MMM and MTA 

It is best to think about the usage of MMM and MTA in the context of planning cycles. It is helpful to think of three marketing planning cycle types: strategic; tactical; and reactive. 

Strategic Planning 

Strategic planning typically happens annually, with more ambitious strategy resets looking out three or even five years. This type of planning typically looks at the total marketing investment envelope (e.g., $50M or $75M per year); the rough mix by funnel position; and any new channels or types of marketing to be tested at scale. MMM—particularly more advanced modeling taking advertising’s impact on brand equity—is the right tool for this exercise. 

MMMs can be extended from measurement to optimization by extrapolating the curves outputted from statistical inference into “what if” scenarios, and then using machine learning to re-mix marketing until an optimal solution is reached. This optimization step can be a helpful input into the strategic planning process. It is important to note, however, that optimizations based on past results are not accurate when predicting huge budget swings. A good rule of thumb is that beyond a 20% increase or decrease in budget, curves become unreliable. 

Tactical Planning 

Tactical planning typically happens quarterly or annually, and looks at the specific channel and audience mix that will drive maximum ROI—sometimes called ROAS in marketing circles (return on advertising spend). Both MMM and MTA can be useful in the tactical planning phase. MMM is good to understand marginal customer acquisition cost (CAC), allowing campaign planners to re-mix channels to maximize effectiveness given a certain budget. MTA can then be mixed in to identify recent changes in channel effectiveness, and to get to granular detail on specific creative types, landing pages, offers, and cadence.  

Reactive Adjustment 

Reactive planning (or adjustment) happens constantly. Marketing dashboards typically start with last-touch results. They become far more powerful when positioned side-by-side with MTA results. Channel managers who are used to seeing last touch CPAs will now see a mutually exclusive, collectively exhaustive view of CPAs, as well, that credits channels with “halo” if they drive more influence up the funnel. MTA is ideally suited for reactive adjustment because it can be built to update in near real time.  

Figure 3: A last touch vs. MTA version of channel contribution. The “MTA effect” is the impact of multi-touch attribution on last touch CPAs. Some channels do better, and others look more expensive. 

Final Thoughts on the MMM vs MTA Debate

In the debate of MMM vs MTA, the answer really is “both”. Both methodologies offer unique benefits and address different aspects of marketing effectiveness. The good news is that they can be built together, using one data pipeline. By understanding their strengths and limitations, marketing leaders can leverage MMM for strategic planning and MTA for tactical optimization and reactive adjustments. Combining both approaches—without throwing away last touch attribution—provides a holistic view of marketing performance, ensuring that every marketing dollar is spent effectively.  

Marketing campaigns typically target one of two main audience pools: 1) Prospects, i.e., those a business aims to acquire; or 2) Members, i.e., those a business seeks to retain or engage. To measure marketing effectiveness, each group presents unique challenges.

In the case of acquisition marketing, information about prospects is often limited. Marketers must allocate spending across various channels to cast a broader net to acquire new customers. They can target addressable markets that meet their specific criteria (i.e., individuals aged 65+ for insurance companies that provide Medicare plans or those that live within specified locations for retail organizations). However, complete information about the prospect pool is often unknowable.

By contrast, member marketing benefits from having extensive data on current members. Much of the person-level data utilized is generally stored within an organization’s first-party data. This allows marketers to utilize personalized channels (such as digital outreach and content-specific mail) for more precise targeting and measurement.

In this blog, we will examine and compare these two types of marketing. For each type, we’ll touch on the broader objectives and measurement techniques, the data required for measurement, the typical challenges marketers face when attempting to gain insights from data, and the ways that analysts can offset these challenges to measure marketing effectiveness.

Why It’s Crucial to Measure Marketing Effectiveness

The measure of marketing effectiveness establishes the incrementality of our marketing, providing clarity into what we gained from a campaign, i.e., would not have happened otherwise. This is essential for discussing budget, ROI, and the net benefits marketing provides to the business. Effective measurement leads to educated, data-driven, strategic marketing decisions that account for unknowns and clarify the ideal focus areas.

Whether the marketing is directed at prospects (acquisition) or members, measurement allows marketers to answer big questions such as:

• “What is the value that marketing is providing?”
• “How can we increase our outcomes given budget constraints?”
• “What marketing tactics are the most effective at driving positive outcomes?”

These are complex questions that drive strategic decisions in organizations. The identification of insights that can provide better answers to these questions and make more informed decisions is invaluable.

Acquisition Marketing Campaign Measurement

Acquisition marketing is complicated. Marketers have a wide range of channels to use to engage new prospects. These include traditional channels, such as Direct Mail and TV (both branded and direct response); the wide array of digital channels, such as Display, Paid Search, Organic Search, Online Video, and Social Media, amongst many others; and more industry-specific channels, such as Events, Affiliates, or Out-of-Home (OOH, e.g. billboards or transit ads).

Measurement Techniques

With so many channel choices, finding the right mix of budget resources can be challenging. Marketing Mix Modeling (MMM) is an excellent method for answering mix-related questions. An MMM is an econometric model that helps measure the effectiveness of marketing activities and their impact on sales or other KPIs. Within MMM model development, audience granularity is not required. A set of aggregated data at various segment levels (e.g., time periods, geographies, product categories, and other types of well-defined metadata) can be used to build a valuable model.

While MMM helps provide a broad understanding of how to allocate spend across channels, it is limited in its ability to provide insights at the more granular customer level. Identifying how to attribute “credit” for each sale (or other output KPI) across marketing channels requires some attribution models.

The default, simpler methodology is last-touch attribution (LTA). This methodology attributes 100% of the sale to the last channel a customer engaged with before converting. This often leads to over-attribution of bottom-of-funnel channels in reporting, causing an overinvestment in these down-funnel channels that is more straightforward to measure; and atrophy within less measurable, upper-funnel channels that play an essential role in the acquisition process through their impact on brand health and audience attitudes.

The far more challenging approach is multi-touch attribution (MTA). MTA aims to solve the LTA problem by providing a more comprehensive view of the customer journey. The model apportions the credit of an application across the various channels that prospects engaged with throughout the buying process. A good MTA model can help marketers better understand the value of each touchpoint. With the complicated ecosystem of marketing and sales efforts, better information can significantly improve strategies.

Data Requirements

One key challenge in measuring acquisition marketing is finding signal within the sheer volume of data available. With a huge market and many potential customers, marketers have access to a wealth of information that can inform their strategies. However, the quality and granularity of the data across channels can vary significantly.

For example, in the case of Direct Mail as an acquisition channel, marketers have complete visibility into the “1s” and “0s” – that is, they know exactly who received a piece of mail and who did not. This level of granularity allows for more precise measurement and analysis.

Alternatively, poor ID resolution for upper- to mid-funnel channels, such as Digital channels and TV, is a crucial hurdle to overcome when building an acquisition MTA. Unlike Direct Mail, which allows for precise tracking, these channels often rely on cookies, device IDs, or other imperfect methods of identification. As a result, these channels usually only yield aggregated impression and response counts. As such, it can be difficult to connect individual users across different touchpoints on these platforms if they don’t use them to convert.

Finally, multiple application capture channels generate different levels of data, which can be challenging to reconcile with each other and the funnel activity. For example, a lead captured through a website form may provide different information than one captured through a call center. Integrating these disparate data sources into a cohesive MTA model can be complex and time-consuming.

Strategies for Overcoming Measurement Obstacles

Marketers can implement several strategies to overcome the challenges associated with poor data granularity within marketing channels. One strategy is an aggregated MTA approach, as opposed to the more traditional discrete MTA detailed above.

An aggregated MTA measures the effect of marketing investments on the last-touch attributed marketing channels or known segments over a given time period. This methodology estimates channel- or segment-specific dependent variables using multiple stimulus variables. In many ways, an aggregated MTA is more similar to an MMM in technical modeling methodology than a discrete MTA. In this instance, it measures the multi-channel impacts on siloed/coded response measures.

Model propensities offer another approach to overcoming the measurement challenges of channels without person-level impression data. This is applicable for channels, such as DRTV or Digital channels, where the individuals who convert (the “1s”) are known, but those who do not convert (the “0s”) are not. The aggregate number of impressions for these channels is known for a specific geography and time period. A model can be built to impute impressions within the broader audience pool. Individuals with higher propensities can then be treated as a “high likelihood” impression within a traditional MTA model.

Member Marketing Campaign Measurement

Member marketing can have many objectives, such as ensuring existing customers remain satisfied with their experience or engaging the individual to take certain actions to add value to the organization (e.g., extending their subscription, buying an add-on, etc.). While having a “known” population makes member marketing campaigns easier from a targeting perspective, these campaigns can face challenges in determining the prioritization of the testing agenda.

Measurement Techniques

A Randomized Control Trial (RCT) is the classic approach for testing incrementality. In this methodology, the population is randomly split into “test” and “control” groups, where the former receives the novel treatment to be evaluated, and the latter gets a comparative baseline. This baseline typically does not include marketing and is considered a universal holdout for establishing marketing’s incrementality in general. This approach allows the team to determine what conversions would have been lost if the campaign had not been run.

While RCT is the most straightforward approach conceptually, there are some opportunity costs to consider. In member marketing campaigns, the impact of traditional RCT universal holdouts tends to be “felt” more by the business because the members being withheld from marketing are known. Additionally, in its purest form, an RCT should ask a simple, one-objective question per test, which, depending on the outcome’s response timeline, can frustrate marketers who want to test and learn on big questions at a faster cadence.

Given the desire to preserve the whole addressable population in most member campaigns, MTA is a measurement route that can provide reads on incrementality while allowing the total population to be messaged. As discussed in our acquisition example, member-level data enables us to apply this modeling concept at a per member, per journey level, or a discrete MTA. The outcome is an understanding of incrementality per conversion while enabling aggregations to multiple different audience cuts given the discrete data origins.

In the member marketing world, a discrete MTA is an attractive foil to RCT, given that it mitigates some of the obstacles noted above. When using this approach, a universal holdout is not necessary to achieve incrementality. However, embedding RCT outcomes into the model itself and comparing the model output against RCT tests is a way to make your discrete MTA more mature.

Since discrete MTA models are built at an individual level and member populations can quickly get into the tens of millions, an advanced data science platform with scalable compute is necessary to enable this type of work (e.g., Databricks or Snowflake). Additionally, temporal constraints on the underlying data feeding these models can sometimes create a barrier to scoring and results. For example, if our campaign’s objective is first to have a member schedule an appointment with their doctor and, after that, schedule another appointment to get a specific test, such as a colonoscopy, the lifecycle of that final response is several months. Additional data complications and refresh cadences can delay this process even further. Given that MTAs look backward on conversions to see what drove the action, the campaigns receiving credit in this scenario would not get proper attribution for several months.

Data Requirements

In a member marketing scenario, we have distinct, discrete data on each person we are marketing to. There is also a clear line of sight into each individual’s outcome, i.e., whether they did or did not act. This means we can cleanly get to the “1s” and “0s,” reflecting the true outcome for a member. This foundation serves RCT and discrete MTA methods, allowing for tailored sampling and individual-level modeling.

Additionally, the marketing typically measured in member campaigns is not usually awareness-based outreach (e.g., TV ads). Instead, it consists mainly of direct response marketing. Channels such as email, outbound calling, SMS, and direct mail can all directly route members to the intended outcome. This facilitates member-level tracking of each marketing piece an individual received and intermediary metrics (opens, clicks) for the applicable channels.

Strategies for Overcoming Measurement Obstacles

Given the tradeoffs between the two outlined approaches, the marketing team should align with stakeholders on a clear learning agenda. If faster results are the priority, RCT would be the best solution despite its tradeoffs. To overcome hesitancy in the control population in the case of a universal holdout, the analytics team can provide various options on how sensitive (what level of confidence) and specific (how small of a difference) the test read will be—the more sensitive and specific, the larger the holdout. Often, a middle ground can be found to produce definitive reads at a reasonable population size.

If there is no desire to withhold individuals from marketing and the conversion metric being tracked has a reasonable lifecycle, a discrete MTA would be the favored method. Once it has been built and given consistent streaming inputs and a relatively swift conversion metric, discrete MTAs can migrate into real-time reporting of conversions reported to date.

Techniques & Strategies for Effective Campaign Measurement

It’s crucial to measure the marketing effectiveness of acquisition and member marketing campaigns to make informed, data-driven decisions that optimize budget allocation and drive positive outcomes. While each marketing method presents challenges, various measurement techniques and strategies can be employed to overcome these obstacles.

For acquisition marketing, a combination of MMM and MTA can provide a comprehensive understanding of how to allocate spend across channels and attribute credit for conversions. In the case of member marketing, RCTs and discrete MTAs can help establish the incrementality of campaigns and identify the most effective tactics for engaging existing customers. By leveraging the appropriate measurement techniques, utilizing the available data, and aligning on clear goals, marketers can make strategic decisions that drive business growth and demonstrate the value of their efforts.

In the fast-paced B2B marketing and sales industry, analytics have become indispensable. Successful go-to-market strategies require a thorough understanding of the five essential jobs that analytics teams perform to unlock success. This blog delves into each job, providing insights and best practices for optimal results.

What are the Five Essential Jobs?

1) The What: Reporting

Definition: The What job is about reporting the facts. How many leads are we generating, how quickly are revenue analytics leads moving through the funnel, how much are we spending on marketing, and how is each rep doing versus their goals? This is the most basic, and also the most important, “table stakes” job. Fortunately, it can be mostly automated via business intelligence tools.

Reporting is the lifeblood of sales and marketing executives, ideally offering near-real-time performance insights. Different types of reports provide insights into various parts of the pipeline.

Types of Reports:

• Activity-Based Reports provide insights into day-to-day actions and engagements that drive a B2B marketing and sales process.
• Performance-Based Reports look beyond activities and assess the results and outcomes of sales and marketing efforts.
• Forecasting Reports anticipate future sales movements and market trends.
• Customer Reports offer insights into customers’ acquisition cost (CAC), their value over time, and the results of targeted marketing efforts.

Reporting Best Practices:

• Ensure data quality for trusted reports.
• Start with a simple report and gradually expand.
• Establish a common data language (taxonomy).
• Address duplicate data issues for accurate reporting.
• Employ centralized data storage for effective reporting.

2) The Why: True Analysis

Definition: The Why job is most akin to the Greek root words of analytics—literally “untying a knot.” Analysts answer never-ending ad hoc questions from executives on almost any topic imaginable. This process cannot be automated; analysts need fast access to clean data and sound data science tools to get results.

Marketing and sales analysts are crucial to the success of a commercial team; they perform the vital task of true analysis, combining business and data science skills to unravel insights and drive informed decision-making.

Key Requirements for Analysts:

• Fast compute and affordable, limitless storage.
• Reproducible analysis through text-based data science languages.
• Timely access to relevant data for analysis.

Agile Project Management for Analysis:

• Use a Kanban approach for task listing and prioritization.
• Maintain a clear list of analysis tasks, stakeholders, and due dates.

3) The Who: Targeting and Segmentation

Definition: The Who job deals with accounts and customers—segmenting them, targeting them, and serving them the right content. This job provides the strategic input for Account-Based Marketing (ABM).

Effective B2B targeting involves understanding market dynamics, including segment targeting, within-segment targeting, and within-account segmentation. The analytics team must integrate various data sources to create target lists and buyer archetypes.

Data Integration for Targeting:

• Segment and Within-Segment Targeting: Industry trends, firmographic data, behavioral data.
• Account Targeting: Master hierarchical table structure for accounts.
• Within Account Targeting: Categorize key players, map decision hierarchies, and use qualitative and quantitative research.

Actionability Requirements for Segmentation:

• Define actionability requirements before segmentation.
• Assign segments back to leads and contacts with reasonable accuracy.

4) The How: Measurement

Definition: The How job is about measuring marketing and sales: How did we get this lead? And what can we do to get more like it? In B2C companies, media mix modeling (MMM) is commonly used to get these answers. This is trickier for B2B companies but just as critical.

B2B marketing measurement poses unique challenges, requiring a nuanced approach. Small-n deals, multiple objectives, chunkier tactics, long-time scales, audience complexity, and sales integration demand a hybrid measurement approach.

Measurement Challenges and Approaches:

• Separate measurement for different objectives (awareness, demand generation, sales enablement).
• Hybrid approaches combining econometric, deterministic, test-based, and heuristic-based methods.
• Leverage management insights to complement data-driven models.

5) What’s Upcoming: Prediction

Definition: The What’s Upcoming job predicts what customers will do or respond to. It is the “action” side of account-based marketing and depends on machine learning techniques (predicting and classifying based on signals).

Machine learning plays a pivotal role in predicting outcomes in B2B go-to-market analytics. However, B2B predictive modeling faces challenges such as sparser signals, fewer observations, and a need for a deeper understanding of individual features.

Challenges and Considerations:

• Signals are sparser in B2B predictive modeling.
• Fewer observations require a more challenging modeling approach.
• Consider outsourcing to ABM platforms but retain a team of predictive data scientists for deeper insights.

Unlocking Success in B2B Go-to-Market Analytics

Mastering the five essential jobs in B2B go-to-market analytics—reporting, true analysis, targeting and segmentation, measurement, and prediction—is key to unlocking success. Staying ahead with innovative approaches and a robust analytics strategy will pave the way for sustained growth as the landscape evolves.

A Data Science-Backed Approach to Brand Building

Any marketer will tell you that brand-building takes significant time, concerted effort, and major resources. Measuring that work, however, can feel like an altogether separate beast. Quantifying the impact of upper funnel efforts on eventual sales is an increasingly high priority for effective marketing teams—those seeking to avoid the measurement trap and balance out their marketing investments.

While Media Mix Models (MMM) are often the best tool for evaluating marketing effectiveness and avoiding last-touch bias, these advanced tools can struggle with the long-term and slowly compounding nature of brand-building. Inevitably, we reach a key question: “How do we take a multi-stage and data science-backed (read: econometric) approach to modeling a brand?”

Figure 1: Generalized multi-stage model

“Brand” as a Collection of Latent Factors

Let’s start with what we know: there is no single “brand equity” KPI that we can observe and track. There are certainly observable indicators of brand strength, things like awareness (be it aided or unaided), NPS, product ratings, repeat customer rates … the list goes on. And, on the other side of the equation, there are observable drivers of brand equity—this is anything that pushes the brand forward (or backward), things like earned media, competitor spend, celebrity partnerships, paid marketing campaigns … again, the list gets long in a hurry.

Once we have collected the relevant data and built a comprehensive list of our drivers and indicators, we can leverage an econometric technique called structural equation modeling to estimate the factor (or factors) in the “middle”—the unobservable thing that is driven by the drivers and indicated by the indicators—a latent factor. With a bit of data science wizardry, we can load these observable metrics against factors and review the strength of those relationships.

Choosing Brand Latent Factors

Again and again, we find that collecting indicators and drivers to load on a single factor (call it “brand equity”) simply doesn’t work. We’ll often generate strange coefficients with low confidence in our modeling outputs when using a single factor. There is no one-size-fits-all approach here, but with modeling iterations, we can start to think about a collection of brand factors that have distinct sets of drivers and indicators. And this is where things get interesting.

When aligning drivers and indicators to factors and selecting how many factors to model, we start a conversation between quantitative data science and qualitative marketing strategy. We collect a comprehensive data inventory, build a multifaceted “story of the brand” using key events and trends over time, and finally let the data tell us which aspects of our story seem to be resonating econometrically. This give and take will help crystallize our mental model and generate a set of latent factors to model on.

Here are some unobservable latent brand factors that might underpin our brand model:

• Brand Awareness
• Category Awareness
• Brand Affinity
• Brand Salience or Fame
• Perceived Quality
• Brand Loyalty

Figure 2: Illustrative map of SEM results for brand latent factors

What Can We Learn from Modeling Brand?

Once we have a model that (a) loads econometrically with high confidence and (b) lines up with our intuition (our “priors”), we can chart these brand factors over time and begin to investigate causal relationships or key moments in a brand’s timeline.

Figure 3: Illustrative brand journey—not all factors move together

As we incorporate brand factors into a holistic marketing effectiveness model by layering on other components (i.e., modeling sales response, controlling for seasonality and macro factors, etc.), we can start to ask the juicy questions that brought us to this exercise in the first place, things like:

• Which brand factor(s) align with the sales trend?
• Which marketing channels drive which brand factors?
• How much of a time lag is there between movements in brand factors and the resulting sales impact?
• How much of a halo effect do our brand factors have on the effectiveness of lower funnel channels, like paid search?
• How do strategic initiatives impact our brand factors, and are we moving the ones that matter?
• In sum, what should we—as a marketing organization—be prioritizing?

Delivering Brand Success

A brand is a tricky concept to put a number on, but doing so is exceptionally powerful in a multi-stage model of marketing effectiveness. Using structural equation modeling with the right drivers, indicators, and latent factors, we can take a data science-backed approach and incorporate a cohesive brand story into a holistic effectiveness model.

Every organization should have marketing effectiveness measurement tools at its disposal. Marketers can’t optimize on last-touch attribution alone, and educating internal stakeholders on how marketing channels work together has long-term benefits.

We’ve outlined four situations when marketing effectiveness tools, such as MMM and MTA, can be particularly helpful.

1. Finance-led Organizations
2. Diminishing Demand Environments
3. Tightening Budget Environments
4. Challenging Brand Environments

1. Finance-led Organizations

When the focus is on expense control, lean staffing models, and profitable revenue growth, marketing is frequently viewed as a cost center. CFOs often push marketers to employ media channels with the lowest cost per acquisition (typically demand capture channels such as paid search and affiliate), which may have a negative long-term impact. MMM and MTA can help finance-led organizations understand what portion of sales are driven by marketing without the risk of overattributing to paid search and affiliates.

Additionally, aligning on marketing-driven cost per acquisition can support (or start) conversations around the level of investment needed to generate additional sales to hit targets.

If finance sets a cost-per-acquisition limit at $100, Display could support an average of $34K daily spend, while CTV could only support an average of $3K.

Or, looking at it another way, if the budget increased by 15%, how many more sales would convert, and how would the overall cost per acquisition change?

Finance may be persuaded that the additional $3.6M in budget is worthwhile to generate an additional 13,000 sales.

2. Diminishing Demand Environments

When an organization has over-rotated toward lower funnel activities, sales demand decreases over time, and cost-pers increase. Marketing effectiveness measurement can help identify opportunities to increase upper funnel spend (to generate demand) through response curves. Response curves show where the next dollar should be allocated to maximize return and when a media will see less return on additional spend.

Once marketing decides to shift investment up the funnel, MMM and MTA can validate that this was the correct move, help track cost-pers by channel over time, and refresh spend curves at higher (or lower) levels to continue to optimize spend.

3. Tightening Budget Environments

No organization has an unlimited budget, and marketers often default to last-touch attribution to determine where to cut when budgets are reduced. Unfortunately, this is short-sighted; last touch significantly underweights less trackable activities, such as TV, online video, and other awareness marketing.

Looking only at last touch in the above example, one would likely cut TV spend. However, the actual marketing ROI of TV is significantly better than what can be measured with last touch. Marketing effectiveness measurement helps make marketers smarter with more information about the performance of marketing channels.

4. Challenging Brand Environments

When organizations are struggling with brand—such as a PR incident, long-run customer service issues, a popular new category entrant, or lack of brand awareness activities—it can be challenging to stay the course on brand spend. Brand-building activities take a long time to pay off, and organizational stakeholders (finance or the board, for example) may want to see proof that investments are working.

MMM can measure the impact of marketing on brand health and, consequently, brand health’s impact on sales. While moving brand metrics significantly takes a long time, modeling can confirm and quantify the link between brand health and sales.

An equation on “responsiveness” and “stickiness” can quantify how much brand spend is required to move the brand metric. “Responsiveness” measures how the brand metric responds to marketing (i.e., how quickly marketing can move the metric). “Stickiness” measures how long the change marketing drove lasts (i.e., how quickly the metric returns to previous levels).

Combining these two elements quantifies how much spend is required to improve a particular brand metric and how long that change will last. Example: One week of $1M brand spend will yield a cumulative increase of 1.7 percentage points in Awareness over eight weeks.

Looking at historical sales performance during periods when brand metrics were higher helps quantify the sales impact of increasing a brand metric calculated in the formula above.

Together, these can help stakeholders outside of marketing understand the sales impact of these activities over a shorter period and buy marketing time for the flywheel of long-run brand improvement to take effect.

Conclusion

Usually, we are brought in when things aren’t going well, and frequently, the issue is one of the four situations covered above. All organizations could benefit from a more holistic approach to marketing measurement, though.

The Value of Zip Code-Based Data in Go-to-Market

Marketing and sales teams rely heavily on data to make informed decisions, drive strategies, and target their efforts effectively. In today’s world, zip code-based geographic and demographic data offer several invaluable insights that can supercharge their endeavors. Different regions can exhibit widely varying consumer behaviors, preferences, and needs. What works in one area may fall flat in another due to cultural, economic, or even regulatory differences.

Consider the Medicare insurance market. In this industry, healthcare policy terms, conditions, and coverages change at the local (zip code) and state levels. Not only do sales and marketing teams face restricted distribution and data collection regulations, but plans and markets are highly competitive. For marketing and sales professionals, the capability to segment by location ensures marketing messages on the benefits of plans are accurate and relevant to the demographic.

Optimizing Marketing’s Impact Using Geographic Data

Many of Marketbridge’s clients are in the thick of geographic-based data every day. The granularity and precision offered by such data can be a game-changer in the competitive landscape, ensuring that efforts are data-driven, precise, and, most importantly, effective. For example, a marketing campaign that resonates strongly in an urban setting may not have the same impact in a rural area. Similarly, a product that sells well in a high-income neighborhood may not perform as well in a less affluent one. By breaking down performance data geographically, organizations can tailor their approaches to align with local conditions, enabling them to allocate resources more effectively and improve ROI.

Predicting Customer Changes

Additionally, a geographic lens on performance data can provide valuable insights into emerging market trends and opportunities that are not immediately obvious when looking at aggregate data. For instance, a sudden spike in sales in a specific region could signal an unmet need or untapped market potential, providing a first-mover advantage to companies that act quickly. Alternatively, a decline in a particular area could be an early warning sign of market saturation, increased competition, or shifting consumer sentiments. Understanding these geographic nuances allows businesses to adapt and pivot their strategies proactively, ensuring they stay ahead in meeting customer needs and expectations.

Examples of Zip Code-Based Use Cases for Modern Go-to-Market

• Targeted Marketing Campaigns: Zip code demographics allow marketers to tailor campaigns, ensuring better audience resonance and higher conversion rates.
• Resource Allocation: Sales teams use zip code data to pinpoint high-potential areas, optimizing resource distribution.
• Personalized Customer Experience: Demographic data informs localized marketing strategies, enhancing brand differentiation.
• Predictive Analysis: Geographic and demographic insights help companies anticipate product or service demand.
• Product Development: Demographics guide product features, pricing, and branding for potential launch areas.
• Optimizing Distribution: Zip code data identifies ideal locations for distribution hubs or outlets.
• Competitive Analysis: Understanding area demographics helps marketers assess competitor footholds and adjust tactics.
• Risk Management: Businesses gauge regional risks using socio-economic data, influencing decisions like insurance premium settings.
• Enhanced Customer Segmentation: Zip code data refines customer targeting, from luxury promotions in affluent areas to budget offerings elsewhere.
• Compliance and Regulation: Zip code demographics help sectors like finance maintain local regulatory compliance.

The Challenges of Geospatial Data Analysis Using Zip Codes

While the significance of geographic data for sales and marketing teams is clear, transforming raw data—whether latitude/longitude, address, zip code, or census block—into actionable insights requires expertise. In the past, businesses leaned heavily on U.S. postal service data for geographic information, often outdated (by over a decade) and plagued with technical inaccuracies. Data associated with zip codes that traverse state boundaries presented additional challenges. U.S. Census Data provides a more accurate alternative data source for mapping, but one-off analysis is time-consuming.

The Roles and Toolsets Needed

Zip code analysis requires the right combination of talent and automated technology. At Marketbridge, “data detectives”—analysts with a deep data science toolkit who explore diverse data sets to understand why events/trends have happened—play a critical role in this endeavor. They use open-source, fully documented code connected to source systems, vendor files, and public geospatial data to standardize analysis and modeling of geographic data changes over time. The solutions these analysts create are document-based with code embedded in workbooks and are easily knittable to reports mapping changes and trends in consumer behavior for sales and marketing teams.

Analysts at Marketbridge use RStudio and the extendable RMarkdown/Quarto framework for this use case. Unlike Jupyter Notebooks, Quarto documents are entirely text-based, making version control easy via Git/Github, and are then convertible into HTML reports, PowerPoint slide decks, or Word documents.

Building an Open-Source Solution for Zip Code Data

Frustrated by the limitations of R mapping when using Census data to overlay demographics on zip codes, I built Choroplethr, an open-source package for the R language, as a solution. Choroplethr seamlessly connects to U.S. Census Data for mapping, providing more accurate geospatial data and automatically handling visual mapping tasks.

Choroplethr is open-source and available to the public; to use it from R, simply type:

install.packages(“choroplethr”)

Choroplethr works very well with a fairly new R package, zctaCrosswalk. This package contains the U.S. Census Bureau’s 2020 ZCTA to County Relationship File, as well as convenience functions to translate between states, counties, and zip code tabulation areas (ZCTAs). You can install the package like this: install.packages(“zctaCrosswalk”).

Mastering Data for Strategic Growth

In today’s data-driven landscape, the nuanced understanding of geographic and demographic data, down to the zip code, is pivotal for marketing and sales strategy. Through tools like open-source software (R studio), professionals can access refined insights that were previously difficult to harness or done via a one-off analysis. These tools not only aid in crafting precise campaigns but also in understanding market dynamics at a granular level. This journey underscores the importance of evolving technology and innovation in making sense of vast data sets.