Incremental Lift: Measuring Marketing Efforts Beyond Attribution
In the pursuit of converting a visitor into a paying customer, there are various channels that marketers adopt for retargeting. As marketing channels have grown further, user journeys became sinuous. Interaction points are multiple and don’t follow any specific sequence. It means the complexities in giving credit to the channels have grown further.
For long, different attribution models have lent some help. Advertisers can choose from different available models (such as the Last touch model, Multiple touch model, First Interaction model, and the list goes on). Based on a predefined model, the credits are shared and strategies are built on them. However, a piece of key information missing in this concept is the direct impact of a channel on the Return on Ad Spend (ROAS).
With the increasing focus at business’ end goals, marketers are looking for metrics that can give deeper campaign insights. And, incremental lift measurement comes into the picture.
In this article, we are discussing the pros and cons of calculating ROAS based on the attribution and incremental lift.
Attribution Models
When you build a campaign that relies on the multi-channel funnel, you expect to retarget the user across different touchpoints. These touchpoints can be third-party ad space, your email, search engine result, or your offline marketing efforts. To acquire some of these touchpoints, you may have pay to your ad partner and other involved parties. The attribution model helps you measure the performance of the touchpoints.
Some of the popular attribution models are Last Interaction, First Interaction, Linear, Time Decay, and Position-Based.[1]
Reasons to Use Attribution Model
Straightforward Calculation
Attribution models follow a predefined and simplistic calculation of the credit for each channel. So, if you are running a retargeting campaign with limited resources, unable to dig deep into the data, the attribution model can be an excellent choice.
Works Even With Limited Data
There is no lower or upper cap needed with attribution models. Regardless of the traffic size, duration, and budget of the campaign, you can follow that calculation and pay it to the respective party.
Several Vendors Follow It
Advertising is a social world, and there are times when you have to follow some practices just because others are doing it. Attribution is a prevalent norm among several ad partners. Therefore, you may have to calculate it as well.
Why You Need to Look Beyond Attribution Model
Assumption Dependency
Attribution models rely on certain assumptions of the value of touchpoints. Some give a maximum share to the last touchpoint, some distribute the share among all touchpoints, and depending on strategy models, it can vary. As marketing analytics move towards a standard where decisions are data-driven and non-generalized, the assumption-based approach loses its value.
More Click-focus, Less Intent-focus
The number of clicks remains fundamental of attribution models while ignoring the click intent of the user. It limits the marketer’s approach in understanding the user behavior and hence, optimizations with the campaign are also cramped.
Ignorance to Immeasurable Contributors
When you run a digital advertising campaign, there are various indirect contributors, including your existing brand value, word of mouth, and offline campaigns. Most attribution models often turn a blind-eye to such contributions to the success of your campaign.
Incremental Lift: An ROAS Focused Solution
Incremental lift measurement is NOT a replacement of the attribution model. However, it gives a more in-depth insight into contributions to campaign performance. Planning ad spend based on just attribution is not enough; incrementality data may provide better spending avenues.
Various channels can contribute to your revenue without getting the user to ‘click’. Let’s consider two examples of Adam, one of your loyal customers.
Example 1: Adam searches your brand name on Google. If you are running Google search ads, Adam sees your paid ad before the organic result. If he clicks the ad, you are charged for it, even though Adam would have reached your website even with the organic result. The ad gets the attribution, but does it add value to your revenue?
Example 2: Adam checks a product on your app and adds it to the cart, but before making the purchase, he is distracted and drifts to a social media app. Now, he sees a retargeting ad of your product on social media, which reminds him to purchase the product. He doesn’t click the ad, but comes back to your app and completes the purchase. Here, if you are relying on the Cost per Click (CPC) ad model, you don’t pay to the ads, even though it played a vital role in a successful conversion.
Incrementality tries to find contributors to the revenue, regardless of click or impressions.
Incrementality measures the lift that a channel brings in the campaign results. It could be for various metrics, including revenue, cost of acquisition, and conversion rate. It helps you figure out the ROI on different (direct and indirect) touchpoints that users interacted with during your campaign before making the transaction. Hence, it is a better indicator of your Return on Ad Spend (ROAS).
Calculating Incremental Lift
In simpler terms, this data-driven approach relies on testing different groups, where some users are not approached over a specific touchpoint, and the difference from the control group is the incremental lift.
If you are already running some campaigns, here are the steps to calculate the incremental profit that you are driving from it:
1. Pick a test group of around 10% of the total audience. So, if your audience is 2000, then the test group is 200. The remaining audience (1800) is your control group, which is equally homogeneous across the user funnel as the target group.
2. For the test group (TG), stop the touchpoint (say banner ad) for which you want to measure incrementality. Continue the usual campaigns for the control group (CG).
3. Run the campaigns like this for an equal duration (say a month).
After a month, let’s assume TG earned $3,000 worth sales and CG earned worth $36,000. Now calculate the average potential revenue of TG for the audience in CG. In our example, TG for 1800 audience size is normalized to $27,000.
4. Incrementality is the difference between the potential revenue of CG and TG, which is $9,000. If your banner ad costs you $10 per user, then total expense for the audience would have been $1,800. So, your incremental profit will be $9,000-$1,800 = $7,200.
The Role of Artificial Intelligence (AI) in Incremental Lift
The practical implementation of incremental measurement involves a deeper level of data modeling that considers several other parameters, like user intent, historical behavior, and more such attributes. When your mobile app (that is serving or planning to serve a million users) will use programmatic ads with different channels, product messages, placements, and formats, incrementality becomes an ‘uncompromisable’ data for you.
When your ads are being served automatically using audience modeling (on several user attributions) to deliver the optimum results, it is near impossible to create and track different target groups manually. A manual effort will be not only time-consuming but also prone to errors. Applying Artificial Intelligence in the incrementality picks and tracks different target groups to automate lift calculation because of different channels.
Why Incremental Lift Is The Way Forward
A programmatic Demand Side Platform, like RevX, provides access to incremental profit for different channels, which enables you to understand how much value a channel is adding to your advertising. Hence, your decisions to optimize your campaign expenses are backed by data and are revenue-focused.
As the advertising industry is stepping towards the ‘gold standards’ of analytics, it is important to gather granular campaign insights. Incrementality can track lift by each channel and even those considered to be immeasurable. So you are rightly equipped to decide your future campaigns as well as optimization with the existing ones.
Summary For Marketers
Attribution and incrementality is not a traditional ‘Android vs. iOS battle’, where you need to pick the only one. These two are calculations based on data from your campaigns, which means that they are both available at your perusal. It is more about what matters for your strategy at the given point of time and which one you should focus on more. Attribution is a helpful metric in tracking the performance of different channels in your campaign, but incrementality utilizes it to take you a step closer to precision in revenue and ROI tracking.
For smaller marketing campaigns that involve counted touchpoints and limited user data, the attribution model can do the job. In case you are running more extensive campaigns that involve deeper audience insights as well as multiple touchpoints, incrementality gets the upper hand. As a marketer who has a clear focus on the revenue or ROAS, you will pick incrementality is a clear winner here.
Reference
[1] Overview of Attribution modeling in MCF — Google
Want to see how incrementality works on RevX — mobile programmatic platform? Book your test campaign now, by reaching us as marketing@affle.com.
Originally published as Incremental Lift: Measuring Marketing Efforts Beyond Attribution at https://revx.io on April 9, 2020.
