This is some text inside of a div block.

Econometrics or Attribution

Back to all articles
Date: June 8, 2020
Category: Blog article
Author: 
Matt Andrew

Fill in the below to receive the Blog article

必填项*
提交成功
点击下方按钮下载白皮书
Oops! Something went wrong while submitting the form.

Econometrics or Attribution: using the core tools of marketing analytics to take brighter decisions.

Granular data doesn’t always lead to better decisions

In today’s data rich world, there is an unprecedented amount of information available to marketers. Tracking customers through impressions, clicks and journeys has opened a variety of new methods for measuring media performance and reaching a deeper understanding of how marketing investment impacts your business. However, with so much information available, having the right tools to interpret this data is essential to making brighter decisions.At Ekimetrics we believe that no single solution can tell every story for every business, and that it is crucial for marketing leaders to have the right tools for the right situation when making investment decisions at the tactical and strategic levels.A key tool for many businesses today is Multi-Touch Attribution (MTA), a process by which brands can track which touchpoints customers are exposed to or engage with on their path to purchase. While this analysis can involve varying degrees of complexity, MTA provides brands the means of directly measuring the success of their media investments by connecting investment with consumer activity at an individual level.However as with all marketing analytics, MTA and attribution models generally are faced with limitations that must be accounted for when interpreting their results. Overemphasis on online media and short-term sales impact can distort the strategic landscape for marketers, too often failing to provide the insights that can effectively steer longer term objectives and planning.In this article we shall investigate the strengths and weaknesses of MTA in contrast with a complementary Marketing Mix Optimization (MMO) approach. MMO is a methodology that leverages longer periods of marketing data with statistical regression modelling to capture the full contribution of media investments to a business over time. Widely implemented as an industry standard for providing strategic marketing insights, leaders across Fortune 500 companies engage with MMO as a cornerstone for long term media planning and driving change across the wider business. MMO avoids some of the key pitfalls of MTA, primarily through its capacity to capture the impact of channels for which attribution is lacking as well as more precisely incorporate baseline, market factors, sales and competitor actions that would not show up in an attribution model.No two businesses or industries are wholly alike, and it is important for any marketing leader to choose the right tool for the right job. In the following sections we shall explore how you can think about what each solution can offer your business and how they can best be deployed. We will then demonstrate the impact that these methodologies have yielded when incorporated into the decision-making process.

Finding Value in differing approaches

To identify the ways in which both MTA and MMO can guide your business, it is key to understand the differences in the methodologies and insights created. Let us consider a representative brand whose business is driven primarily by online sales through a brand owned website platform. This business model affords the brand access to a high volume of user level data at the point of purchase and the ability to track the source of the incoming traffic to the website. The brand’s media strategy however involves a mix of online and offline media, for which they have varying levels of granularity.Such brands can typically make good use of a Multi-Touch Attribution approach to track the day to day touchpoints and take tactical decisions based on customer engagement. However, if we consider how different attribution models assess the impact of media investments, we can see the shortfalls in the methodology and the opportunity to leverage alternative analyses.In our representative scenario, the brand runs a campaign that invests 60% of its budget into online media (SEA, Social Media, VOD, etc.) and 40% in offline media (TV, Radio, etc.). Furthermore, the brand has a high availability of granular user data for 80% of their sales which come through the brand’s website. When the brand tracks user purchases based on the last-click of the purchase journey, they can track 30% of all sales coming through direct and other organic channels (Base) with 50% of sales associated with online media investments.In this scenario, MTA can offer vital insights into the success of different parts of the brand’s online campaigns, attributing value to campaigns that may not have shown up as heavily in last-click metrics. For instance a social campaign that failed to drive a high volume of clicks may well have generated a lot of other engagement among those who were exposed, only later to arrive at the point of purchase through organic traffic or paid search etc. MTA can help brands identify these successes and steer the day-to-day investments in online media.Even in this simplified example, we can find clear opportunities for marketers to leverage great impact from MTA insights. However, when trying to make broad evaluations of media impact and brand health across channels, marketing leaders will run into the challenge of how to properly account for the impact of channels for which only aggregated impact data exists or for external factors beyond the brand’s own media activities: for instance if attribution and budgets are unchanged, how could a marketing team investigate the cause of a decline in sales?MMO enters here with the aim to provide our brand with a full funnel view of their sales and marketing activities. By taking advantage of statistical regression techniques, brands can build a holistic view of the full range of factors that drive their business. In our scenario, the implementation of MMO yields a clearer estimation of how offline investments contributed to the brand performance while also evaluating how concurrent investments from different competitors dampened the impact of brand efforts.

Using MMO to guide your business strategy

The historic nature of MMO modelling makes it well suited for developing marketing strategy and steering budgets for the mid to long term. Marketing leaders tend to lean more towards MMO for guidance here and can generally extract value from those findings over four key objectives:

  • Optimizing your global budgetMMO, as the name suggests, optimizes your overall mix and allows for better allocation of marketing funds
  • Providing key tactical support to campaign implementationWith increased data granularity, MMO can show under which context campaigns and strategy rollouts should be executed to achieve maximum impact
  • Modeling the marketBy including a wealth of market variables, MMO can test out various marketing scenarios in a safe environment
  • Detecting market trendsThe ability to examine a long timeframe of data allows this method to detect overarching trends

Using MTA to guide your business strategy

While MMO can provide insights into the longer-term performance of media, they do fall short when trying to explain emerging trends or the shorter-term fluctuations in marketing performance. Here MTA is implemented to enable businesses to deep dive their digital channels for achieving the following objectives:

  • Assigning value to each touchpointThis is the core of MTA that allows users to quantify the contribution of each stop along a customer’s journey
  • Understanding customer behaviorFollowing each customer’s footprint translates into identifying patterns in behavior enabling key insights to be drawn
  • Ensuring business agilityThe shorter time frame (typically 1 to 4 weeks) means that the clients are able to react quickly to any sudden changes
  • Delivering vital support for product launchCombining all the above elements allows for continuous and rapid improvements to be made via reinforcement learning feedback loops

What’s right for your business?

With limited budgets it is important to be able to take an informed decision between these two methods that best fits the strategic and financial needs of your business. The first level to consider is whether your data is concentrated in online or offline channels. Online only platforms, especially relatively young ones, should invest in producing user level data if possible, to make use of MTA’s quick feedback cycles that can be used to guide marketing decisions early into an analysis window.For those with a mix across online and offline channels, a second important question arises about budget, and those with substantial investments should lean towards a unified approach that begins with MMO and identifies key areas where further insights can be extracted by using MTA as a digital deep-dive tool. Through this unified approach, clients can get a more refined understanding of their ROIs and recommendations that drive long-term success and not just short-term sales.While the rest, with a more limited budget, are better off focusing their investments on MMO because of its more holistic nature that enables it to cover all aspects of the business and its environment.

Conclusion

With analytic methodologies constantly evolving and improving, choosing the right tool for measuring marketing performance is essential for making brighter investment decisions. No single solution can provide a ‘silver-bullet’ to unlocking superior marketing performance and in many cases marketers find themselves reliant on an inappropriate or incomplete vision of their business, putting them at a disadvantage against competitors. Understanding how your media strategy can benefit from both focused shorter-term results of MTA and the broader strategic insights of MMO can help you build a unified approach that will lay a firm foundation for deeper analysis and greater agility in marketing and sales.---

Get in touch

Connect with our Data Science experts

必填项*
谢谢!
We will be in touch very soon!
Oops! Something went wrong while submitting the form.