Marketing Attribution Software 2020: Ultimate Guide

Marketing Attribution is the process of evaluating how important each customer interaction touchpoint is in nudging the customer along towards the desired outcome (“conversion”). Over the last decade, digital communication platforms have complicated marketers’ tasks exponentially. Customer interactions now span a complex range of behavior across a wide variety of media and devices. Not only do customers interact with a company/brand in far many more ways than before, but they also do so in far more complex ways. This complicates marketing decisions around the medium, content, and timing of advertising, sales promotions, POS, and merchandising support, among a whole host of customer interaction situations.

Thus, the ability to make optimal media-spend decisions that “convert” customers to desired outcomes now depends heavily on the ability to track and analyze their behavior across a variety of touchpoints, in various combinations of sequence and frequency.

This complex evaluation process of marketing touchpoints is enabled by software solutions, which run any of several Attribution Models, ranging from very simple Single Source Attribution through to complex Algorithmic ones. Marketing Attribution enables today’s marketer to evaluate the effectiveness of marketing touchpoints at, both, the granular level of the individual customer’s behavior, and at the aggregate level simultaneously, and in real-time. This arms the marketer to adjust and pivot her marketing programs as necessary to deliver ROI, in terms of conversions, that is significantly higher than would have been possible otherwise.

Of the several Marketing Attribution models available (and more are sure to be developed in time), picking the one(s) to be used by a particular business has to be a well-considered decision. One size does not fit all.

For example, the B2B buying process is complex and commonly involves multiple individuals and groups weighing in on the decision. So, in general, single-touch attribution models would be ill-placed to map relevant attribution scores for B2B cases. Yet, if the objective of the attribution exercise is, say, to assess the best method to boost brand/product awareness in B2B audiences which leads to conversion, the first-touch attribution could provide quick and accurate insights.

In some B2C situations, Lead Creation is not a single, distinctly discernible stage, where, using a Lead Conversion Touch attribution model may be less than useful. Thus, selecting the right model(s) requires careful discernment, which businesses typically can use some help in.

Software is what makes effective Marketing Attribution possible.

Historically, marketing budget allocation depended on regression analysis of aggregate data, past practices in the company, and the individual marketer’s hunch as to which channels and methods warranted receiving what part of the pie. The last decade’s explosion of digital social media has put paid to the effectiveness of such rough heuristics. The process of customer conversion has become complex to such a degree that data-driven Marketing Attribution analysis is fast becoming the default ‘correct’ method to strategize for marketing and to allocate budgets.

Attribution Software can run models as simple or complex as required in individual business situations. They can analyze large, complex datasets of customer behavior, throw colorful light on the effectiveness of particular marketing interventions, and thereby enable quick decision-making.

Source: https://www.predictiveanalyticstoday.com/top-marketing-attribution-software/

Marketing Attribution Software leverages advanced (data science) and machine learning to objectively determine the impact of each marketing touch(point) along a customer’s journey toward conversion. – “Top 22 Marketing Attribution Software”, Predictive Analytics Today

Classified by the type of modeling logic used, there are three broad categories of Marketing Attribution Software.

Single-touch and Multi-touch Attribution Software (ST- and MTA)

Multi-touch Attribution Software is used to identify and assign weights to marketing touchpoints across multiple channels to establish their sequence and influence on customer conversions. Multi-touch Attribution helps determine the value of each touchpoint in driving conversion.

Several attribution models, as described below, are commonly used in multi-touch attribution software. Two of the models (the first two below) attribute conversion to a single touchpoint as solely responsible, and software using these are called Single-touch Attribution Software.

First-Touch (or First Interaction) Attribution

Under this model, the first traceable interaction that leads to the customer finally converting is considered as responsible for the conversion. This is useful in cases where a series of marketing touchpoints is planned, in which the success of the first touchpoint is critical to the completion of further touchpoints and eventual conversion.

Last-Touch (or Last Interaction) Attribution

This model lays conversion success entirely at the door of the last touchpoint. This is the simplest attribution model, as it does not require tracking of touchpoints prior to the last one before conversion.

Linear Attribution

The linear model apportions credit equally to all the touchpoints that the customer goes through prior to conversion. The premise of this model is that every touchpoint has equal value in bringing about the desired outcome.

U-shaped Attribution

As the name suggests, this model considers the first touchpoint in the customer’s interaction journey and lead conversion as the most important, while also crediting the in-between touchpoints, albeit to a lower degree. Typically, the first touchpoint and lead conversion are attributed to generate 40% each of total conversion value, with the remaining 20% value brought in by the intermediate touchpoints.

W-shaped Attribution

This is similar to the U-shaped model in crediting the first touchpoint and lead conversion with higher importance, but differs in assigning equal value to opportunity creation. Thus, usually, 30% value each is assigned to the three touch points considered important, and the 10% that remains goes to the remaining touchpoints.

Time Decay Attribution

The Time Decay Model holds that touchpoints are as valuable as they are close to the conversion point. Thus, the last touchpoint is assigned the highest value, value decreases with distance of the touchpoint from the end. The first touchpoint receives the least, usually non-zero, weightage.

Full Path Attribution

This is a highly technical attribution model that is suitable for organizations that do marketing primarily to existing sales prospects. It is similar in treatment to the W-shaped model, but adds a fourth touchpoint to the list of high-value ones – the customer close. 90% of attribution value is assigned equally to these 4 touchpoints, and the leftover 10% is distributed among any remaining touchpoints.

Custom/ Algorithmic Attribution

In cases where the standard models do not accurately reflect the customer’s touchpoint path toward conversion, a custom attribution model may be built, incorporating an algorithm that reflects the real conversion process.

Attribution models aren’t perfect, of course. Multi-touch attribution models do use approximations in the assumptions they make, and so, may not exactly correspond to every customer’s actual personal experience. Certain restrictions imposed by large platforms like Google, Facebook, and Amazon, and some browsers, also make it difficult to access relevant data. Besides, offline experiences or touchpoints are off-limits for these models.

Marketing Mix Modeling (MMM)

Unlike Single- and Multi-touch attribution models, Marketing (or media) Mix Modeling does not track and analyze individual touchpoints of each customer’s conversion path. Instead, Marketing Mix Modeling uses multivariate regressions on aggregate data to calculate the impact of specific sales and marketing activities on customer behavior.

Data pertaining to environmental factors, such as seasonality, competition, macroeconomic conditions, influencing touchpoint outcomes and final conversion, which are not considered in multi-touch attribution models, find a place in Marketing Mix Modeling.

Multi-channel Attribution (MCA)

Multi-channel attribution uses individual level data, but its focus is on weighing attribution credit by channel and not per touchpoint. In some versions of MCA, data pertaining to online and offline channels are incorporated to map out the customer’s full path to conversion. Unified models (called Unified marketing Impact Analysis or Unified Marketing Measurement – UMM) that blend MTA and MMM seamlessly into a single set of rules are becoming increasingly popular.

The benefits of UMM over MTA and MMM are:

Online and Offline data:

UMM delves into data on offline aspects of campaigns alongside the online ones, thus providing insights into customer behavior across disparate online and offline channels in a normalized, integrated way that is more reflective of the real world.

Aggregate and Touchpoint-level data:

Trends and environmental factors are represented by aggregate data, which provide the context for touchpoint data, together presenting customer conversion paths in their real context.

Real-time insights:

Advanced UMM platforms can provide real-time insights and analysis, allowing marketing to pivot mid-campaign to optimize spend.

Historically, marketing teams have always attempted to determine the relative value of individual advertising, promotions, and other customer interaction touchpoints using imprecise, indirect methods. These exercises, which may be characterized as rudimentary attribution, were conducted by analyzing aggregate data and by using proxy data for customer touchpoints.

Modern Marketing Attribution exists only because Marketing Attribution Software exists. Marketing Attribution relies entirely on the capabilities of modern Software for comprehensive, in-depth results. As well as this existential benefit, Marketing Attribution Software delivers the following pluses too:

• ROI increasingly drives the operational details of marketing campaigns

• Reporting across methods/media can be brought into a single-view dashboard, as opposed to standing in completely detached silos earlier

• Competitive intelligence pertaining to performance across different marketing delivery media can be more readily accessed

• Data – and insights – can be viewed at macro and aggregate levels just as much as drilled-down to individual users’ data

• Even disengaging behavior like uninstalling can be used to analyze the best course for advertising strategy and budgets

• Determining the attribution weight of a touchpoint can go hand-in-hand with estimating potential revenue from successful conversions at that touchpoint

• Insights obtained into customer behavior through the attribution exercise can be used to further fine-tune the weightages and models underlying the attribution exercise itself, enabling recursive improvement in prediction accuracy

Tracking is an important aspect of any attribution. Marketing Attribution software products track Visitors, Campaigns, Site search, Pageview , User interaction, Cross-device use, Events, Conversions, Ad spend, so on.

AI in Marketing Attribution Software

Marketing Attribution is a highly dynamic and adaptive exercise at its core. AI forms the basis for much of the data-crunching and analytics that enables attribution. Starting with data collection across devices and channels, profile synthesis, analysis of touchpoints across the whole breadth and depth of the conversion funnel, through to predictive analysis of likely future customer behavior - both long-term and instantaneous and insights and channel-specific recommendation engines, all of this is AI-enabled.

According to the Digital Marketing Magazine, AI-enabled Marketing Attribution significantly enables Marketing and Sales to work more seamlessly together to enhance the customer conversion experience:

“AI platforms being implemented will give the marketers a better view of campaigns allowing them to see if they were successful or not and shed light on whether there were any breakdowns in the campaign. AI will also allow the marketers to see whether the sales team has followed up leads, eventually aligning the discrepancies between the two departments, allowing for better communication.” - (from Accurate Attribution: How can AI make a difference?)

As we have seen above, Marketing Attribution is a potentially complex, yet business-critical, piece of software. Choosing the right Attribution Software involves some key parameters for consideration.

Device/Channel breadth

The number and variety of channels that the Software can include to source data, ranging, say, from mobile-only to full-spectrum online plus offline channels. The range of devices/channels that your business needs require must be supported competently by the tool.

Attribution Models Supported

Your business needs should drive the variety and complexity of attribution models that you should want supported by the Software you purchase. If your organization’s business reality requires the ability to test, optimize, iterate, or pivot attribution models, the Software must present a robust capability to support such changes.

Flexibility and Accuracy of Analysis

Must have the flexibility to support analysis and insights accurately at aggregate level just as well as for drilled-down touchpoint level for the kind of attribution model you want supported.

Reporting flexibility

Especially in multi-channel attribution, the software must have the flexibility to report on each channel separately, as also present a unified view of all channels simultaneously. It must use intuitive reports and dashboards to do so.

Ease of implementation

Implementing Marketing Attribution Software in an organization, due to the multitude of sources that it pulls in data from, is necessarily heavy on integration. Whether this translates into cost-intensive integration effort or relatively quick plug-and-play in your specific organizational context is a key factor in the purchase decision. Depending on the breadth of channels you want to incorporate into the attribution exercise, the attribution models you want supported, reporting accuracy and flexibility, and pricing, you will want to consider current and future integration requirements while drawing up your consideration list of Marketing Attribution Software.

ROI from the Use of Marketing Attribution Software

Although likely exaggerated even in its time, this century-old quote sums up the conundrum that Marketing Attribution solves, viz., getting better and better at getting a louder and louder bang from the proverbial marketing buck.

Marketing Attribution is about measuring, and thereby improving, the Return on Marketing Investment. When selected intelligently, Marketing Attribution Software can deliver directly to the bottom-line by enabling you to enhance the quality of your marketing campaigns more and more.

Attribution insights at the node level can be used to make micro-level adjustments in campaigns.

Macro-level insights too are available, and can be used to pivot and entire campaign segments.

Aggregate-level insights enable Marketing Mix Modeling and refinement in a unified manner across digital and non-digital platforms, over relatively longer periods. Entire multichannel campaigns and even multiple campaigns can become much sharper focused on a continual basis, if Marketing Attribution insights are acted upon.

What are the Challenges in using Marketing Attribution Software?

When done right, implementing Marketing Attribution pays handsomely. The challenge is in getting it right. Choosing the right combination of attribution models to be applied to the right touchpoints is an involved, continuous process.

Paralysis

Getting Marketing Attribution right is usually an incremental process. The team has to be at it, examining results, tweaking and pivoting models, and never putting off follow-up action based on the insights generated. This level of commitment calls for focus, dedicated resourcing, and likelier than not, executive championship of the program.

Very commonly, the paralysis in progress is due to fatigue from trying to achieve perfection early on. As with any other worthwhile endeavor, the key is to simply begin. Then, going forward from there involves staying committed to implementing actions indicated by the previous run of the model, evaluating the results of such action, making modifications in the model, then running it again. Iterating this process is what delivers more and more efficiently effective conversion results from your marketing campaigns.

Judgement Errors

Having overcome extreme reluctance to begin, we could be sailing along and yet subjecting ourselves to incorrect, irrelevant analyses due to any of several possible judgement errors. Some of them are laid out below.

False correlation

A conversion event may get associated with a touchpoint with significant frequency, and yet, the correlation may be coincidental. The touchpoint may not actually have a causal effect on conversion, but is still considered a cause.

Misjudged pre-decision

A pre-decided customer (or one who has already made her decision based on an untracked touchpoint) converts, and the system attributes it to the latest campaign.

Disregarded content effect

Content varies significantly across platforms, yet, the weightage assigned gets determined entirely by differences in behavior on those platforms, ignoring the effect of the content itself.

Incorrect ROI attribution

ROI attribution suffers from a lack of nuance and does not reflect the true effect of the touchpoint.

Failure to conduct Unified assessment

Missing contributions to the conversion effect from offline touchpoints simply because the corresponding data is not being captured and is unavailable.

Disregarded context of conversion

The conversion event may exist in – and be influenced by – relevant factors in its context, but which have not been taken into account in the adopted model. The effect of the brand, regulatory environment, etc. could be some such contextual factors.