Accurately analyzing marketing effectiveness is essential to improving performance and growing business. However, privacy policies like the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) have posed challenges to the collection of user-level data, and marketers have been pushed to find ways to measure without violating user privacy.
The last-touch attribution (LTA) model, which has always been a trusted and popular choice, also adds confusion. For sure, its simplicity and cost efficiency is unbeatable, but there’s a possibility of the model not telling the whole story. Top 10 Forbes marketer Neil Patel even dared to say that LTA is telling blatant lies directly to your face.
With the marketing and advertising environment rapidly changing, it is crucial for marketers to choose a mobile measurement partner (MMP) that puts itself in the shoes of marketers. Hence, in this blog post, we’ll be discussing the steps Airbridge has taken to address marketers’ concerns and take attribution to the next level.
LTA has been the standard option for measuring marketing performance because it simplifies the complicated process of attribution analysis. The model tracks with great certainty the channel that serves as the last touchpoint, making it easier to optimize campaigns and creatives. This shows why most MMPs such as Airbridge use LTA.
However, LTA has its own limitations in that it gives 100% of the credit for a conversion to the single last touchpoint in a conversion path and does not dig into earlier touchpoints that could have had an impact on the user. In fact, Airbridge data revealed that at least 30% of conversions occur after two or more touchpoints. Moreover, given users’ latent inclination to convert, there is a chance that the impact of the last touchpoint could be overestimated.
Airbridge believes multi-touch attribution (MTA) and incrementality measurement could be the solution. The MTA analysis allows marketers to accurately see how each touchpoint contributes to conversion and the Airbridge incrementality model helps measure the true effectiveness of marketing in influencing user conversion.
Airbridge uses Propensity Score Matching (PSM) as well as other statistical methodologies to determine marketing incrementality. The Incrementality Report, updated every day, shows the incremental growth derived from each marketing channel, facilitating media planning and efficient budget allocation. Furthermore, by adopting an observational method rather than an experimental method, Airbridge provides a more flexible view of the incremental effectiveness of marketing campaigns.
User-level mobile attribution is becoming increasingly difficult due to global rise in data privacy regulations. In response, Airbridge proposes Marketing Mix Modeling (MMM) as an alternative to measure marketing performance without collecting user-level data.
MMM uses aggregated historical data to explain the relationships between marketing activities and conversions. There are two primary approaches: Frequentist and Bayesian. MMM came on the scene decades ago when companies with big ad spends were looking for ways to determine the right media mix. Back then, MMM solutions were not easily accessible or affordable because they were long-term high-cost consulting projects, requiring at least six months to one year’s worth of data.
However, MMM is making a comeback as the use of traditional marketing analytics has been limited by its reliance on user data. Combined with LTA and MTA, MMM, which is an aggregated data form of modeling, can effectively attribute iOS campaigns even in iOS 14.5+. It can also be used to measure the performance of Connected TV (CTV) and Digital Out-Of-Home (DOOH) ads.
Another reason why MMM is on the rise is that it can make predictions and run simulations. For instance, you could get recommendations on how to reallocate your ad spends to maximize sales moving forward. Mobile attribution, technically, is based on a retrospective approach, but MMM could take it to the next level, prescriptively determining the best course of action for a given situation. For such reasons, there are B2B SaaS companies trying to lower the bar and let more marketers use MMM with ease.
Airbridge, as a proud participant of Meta’ts MMM SaaS Incubator Program, has been developing various MMM features using the Frequentist approach. Powered by advanced machine learning, Airbridge’s proprietary model measures the performance of each marketing channel and suggests ways to maximize budget efficiency. The Marketing Mix Analytics Report will show how the model has been trained and the Budget Optimization Report will tell how to best allocate within a fixed budget and maximize performance.
Nonetheless, MMM does have limitations because it estimates marketing impact on historical business outcomes and even makes projections based on probability. Airbridge therefore plans to improve accuracy through regular calibration.
In an effort to relieve marketers’ concerns over the limitations of LTA and tectonic shifts in the marketing industry, Airbridge has been proactively investing in research and development. We’re building the Unified Measurement Stack which combines LTA, MTA, and MMM to be able to measure incremental effectiveness of marketing activities without solely relying on user-level data. Our ultimate goal is to capitalize on the strengths of each approach and provide a balanced view.
Airbridge aims to take a step further, striving to find what marketers’ worries are and how to solve them. We believe that our job does not end by simply selling our product, but that we have to guide marketers through the world of marketing analytics, contributing to the growth of their businesses.
So far, we’ve explored the journey Airbridge has taken to establish Unified Measurement Stack. In fact, marketers are often confused using LTA, MTA and MMM because they are based on different data sets and models, meaning that they will most likely draw different conclusions. Moreover, being able to choose among various approaches could only add confusion about what to use when. Hence, in this section, we’ll discuss how you can best use these three solutions.
What is best for your product depends on its characteristics and the market environment, and you need to build a tailor-made Unified Measurement Stack. However, rather than jumping to the conclusion that there is no single answer, Airbridge would like to take responsibility as a solution provider and tackle challenges together with marketers.
Below is our suggestion on how to use each solution to adopt the full-cycle marketing strategy along the user journey. Continue to read to imagine your own Unified Measurement Stack.
MMM stands out in that it uses a time series regression model based on various types of aggregated data. Therefore, it is more suitable for figuring out overall trends or general projections than drawing specific insights. In addition, the aggregated data allows to measure performance of offline channels such as TV ads and magazines, from which it is impossible to collect impressions or click data.
Considering such characteristics, MMM can be best used for marketing strategy development. Again, MMM reads the trends and suggests directions by evaluating both digital and non-digital channels that could drive growth. If you’re using Airbridge, you’ll be able to benefit from the MMM-based Budget Optimization Report to find out your most effective budget allocation.
With Airbridge’s incrementality model, you can discover the true incremental effects of each marketing channel you’re using. In other words, this is a great way to attribute multi-channel campaigns than to see through general trends.
This is why we recommend using MTA and incrementality for measuring and monitoring the results of your marketing campaigns. You can reveal channels that provide incremental growth without heavily relying on the cost calculated by ad networks. Identifying the most effective channels and allocating more budget to them can help improve marketing performance.
LTA is the most commonly used and the most granular way to measure marketing performance as it is based on raw data. Hence, this solution comes in handy when you’re adjusting your marketing strategy to optimize campaigns and creatives.
LTA might not be the ideal option to check the overall efficacy across various marketing channels. However, it could be useful when you are changing the details of your strategy after distinguishing the most effective channels and creatives with MTA. In Airbridge, for instance, you can refer to the Actuals Report to see the performance of each campaign and creative and draw insights for your next campaign.
As emphasized earlier, Airbridge hopes to build the Unified Measurement Stack relevant to marketers’ day-to-day responsibilities. We try to think from marketers’ perspective and go the extra mile to be a reliable partner through the different stages of full-cycle marketing.
As a marketer-centric solution provider, Airbridge prioritizes the needs and wants of marketers. Thus, our MTA and MMM solutions, two of the three pillars of the Unified Measurement Stack, have been developed with some unique features introduced below.
Incrementality can be measured using either experimental or observational methods. Meta and Google’s lift studies are based on the experimental method. Users are randomly divided into a test group and a control group, and the test group sees the ads whereas the control group does not. While this guarantees accuracy, it takes a relatively long time and incurs financial costs.
On the other hand, Airbridge has adopted the observational method, meaning that we use machine learning and statistical methodologies to calculate incrementality. This requires shorter time and lower cost than conducting actual experiments, and even more, marketers can get a daily update on the incremental effect of their campaigns.
In addition, every three to six months, Airbridge compares measurement values delivered by its observational model with those of a calibration standard of known accuracy. Korean health brand Dano ran studies and found that Airbridge’s MTA study results were significantly higher than the LTA results, which better aligned with its actual business results. Compared to Meta’s lift study, the MTA results were only slightly lower. This gave Dano further confidence in its campaign and the company is still working with Airbridge to gain accurate insights.
MMM is great to be used at the strategy development stage of full-cycle marketing. However, since it relies on historical data to determine the average of various marketing results, it can reflect recent changes only to a limited degree. The problem is that today, the marketing landscape is always shifting and that marketers have to respond immediately, coming up with effective strategies.
Fortunately, Airbridge’s MMM solution is more responsive and flexible than traditional ones. Our model, powered by advanced machine learning, can be fully trained with only three months worth of data. We provide daily updates so that marketers can better understand how different channels have contributed to their targets and adjust their media mix accordingly. Moreover, going further from retrospective observation, Airbridge’s MMM proactively and prescriptively recommends media mix to maximize performance.
Many solutions have been trying to overcome the limitations of LTA and deal with the privacy-first era. With more choices, marketers are asking for a unified approach to measure performance more efficiently and effectively.
Airbridge believes that our role does not end at simply providing solutions, but rather, goes beyond building the Unified Measurement Stack with LTA, MTA and MMM and making sure that it adapts to the shifting expectations of marketers. We’re not only a measurement partner but also a marketing partner, and will continue to share various success cases and profound insights.