What is Incrementality? (And How to Measure It)
In theory, assessing the true impact of your marketing, by establishing the true level of marketing incrementality associated with specific activities, should be the number one priority for all marketers.
But, in practice, many marketers are struggling to get to grips with this due to a number of issues ranging from siloed data analytics and ‘walled gardens’ to limitations in existing attribution approaches that only scratch the surface.
So, what exactly is incrementality and why does it matter? And how do you go about measuring it effectively?
We take a closer look at all of this below.
- What is incrementality?
- An example of incrementality
- 4 reasons why measuring incrementality effectively matters
- How to calculate incrementality
What is incrementality?
Incrementality is a way to accurately measure the impact a specific marketing activity (e.g. seeing an advert for a lawnmower) has on a desired outcome (e.g. purchasing a lawnmower). But in such a way that it removes the inherent consumer demand that already exists for the product category and brand.
It measures true lift in your campaigns by effectively identifying sales that wouldn’t have happened in the absence of a specific interaction. Like seeing the lawnmower advert above.
It is a concept that can be applied to measure the true value of a range of desired marketing outcomes including awareness, web visits, conversions, revenue and profitability. And, done accurately, can provide a true picture of marketing performance at a channel, campaign or even ad creative level.
An example of incrementality
Let’s take an example to provide a bit of practical context on the concept of incrementality.
- You are in the market for a new lawnmower and head over to Amazon to review your options.
- In the course of your web-browsing you click through an ad for one of the better-known brands – Bosch – which takes you into an evaluation of your options. This includes listings for better known and trusted brands like Flymo and Mountfield but also less well-known options like Challenge.
- Whilst the Challenge mower range is cheaper – and the reviews for the product are good – the reality is that you have a limited relationship with the brand. So, you decide to stick in tried and trusted territory with one of the more established brands.
- And you decide to buy a Bosch.
Now for a key attribution question: should the marketing team at Bosch attribute the value of the sale to the Amazon ad you clicked through initially?
Definitely not in this case.
Your decision to buy Bosch, and ignore what looks a reasonable alternative from Challenge, is based on your trust in the Bosch brand. And that is likely to be down to a range of marketing interactions. From TV and digital advertising, to instore exposure to product or a personal recommendation from a friend.
This is the key challenge for marketers in unravelling that complexity right now.
For example, how much is their investment in TV advertising for Bosch lawnmowers really influencing purchase decisions – and what is the incremental difference to lawnmower sales of increasing or decreasing ad spend levels in a range of different media channels and campaigns?
That is where effective incrementality comes into its own.
4 reasons why measuring incrementality effectively matters
We think there are a number of key reasons why getting to grips with incrementality is central to an effective marketing approach.
1. Provides a single source of the truth
One of the key challenges for marketers right now is how to get an accurate view of the truth in measurement terms for their marketing activity.
Not only are consumers undertaking increasingly complex journeys across a broad range of channels as they research, evaluate and purchase products and services. But marketers are also increasingly frustrated in their efforts to get a clear analytics view of these customer journeys with 60% of marketers indicating that they think that data to support cross-channel decision-making is broken.
The issue has its root cause in a number of challenges including the siloed nature of the analysis process with teams from brand, organic, paid and programmatic effectively ‘marking their own’ homework. But with more and more ad spend pouring into paid social channels like Facebook, which takes a ‘walled garden’ approach to measurement and data sharing, there can be real issues of trust around the data that the ad platforms are putting in front of you.
Take the example below where Google Analytics (GA 360) is showing attribution at a fraction of the level that Facebook is reporting.
2. Gives you the data need to scale success across your marketing mix
Identifying true incrementality is key to the overall success of your marketing efforts.
It does this by providing you with the data-driven type of insights you need to improve your marketing strategy across the board including:
- Identifying the best overall marketing mix for your business
- Proactively optimising your individual marketing initiatives by uncovering the most effective – and cost-effective approaches – at a channel, campaign and creative level
It also enables you to take a proactive approach to modelling potential changes in your mix and answer key questions like:
- Which media channels and campaigns are really contributing to revenue and ROI?
- What will happen if I transfer non-effective spend from Channel X into Channel Y?
- What impact will launching new campaigns have on conversions and revenue – or do I run the risk of cannibalising existing spend?
And crucially, it gives you the type of data needed to shorten the timing between key media interactions and decisions to buy.
3. Enables you to justify budget allocation by connecting spend to revenue
Being able to connect that marketing budget you spend with revenue that the business generates is the holy grail for marketers.
However, many marketers struggle to make the connection and our own research points to the fact that almost 68% of Marketing Directors report that internal stakeholder pressure actively restricts the option to employ marketing activity with longer term payback.
Effective measurement of incrementality enables you to step out from the shackles of this type of internal pressure. By enabling you to draw a straight line between the investment you make in your marketing activity and revenue generated.
4. Futureproof your marketing measurement at a time of increasing levels privacy
The reality is that the marketing landscape in the process of a significant overhaul.
As the impact of the death of third-party cookies and iOS14 affect marketers’ ability to effectively target and analyse the effectiveness of campaigns the issue of incrementality is becoming even more pressing.
How to calculate incrementality
So how do you effectively measure incrementality in your marketing activity?
The simple truth is that more traditional forms of attribution fall short in terms of their ability to measure incrementality effectively. These include:
- Single touch attribution models
Also known as First-Click and Last-Click attribution, both of these models are a gross oversimplification of what is actually happening on the ground.
They attribute 100% of the credit to the last interaction in a customer journey. So, if we take the example of someone buying a laptop where they have researched the purchase on review sites (and clicked through to product content on the manufacturer sites), seen TV ads for the product, taken recommendations from work colleagues and finally clicked through on a paid Ad on Amazon to buy the product.
In a Last-Click model all of the credit goes to the paid Ad on Amazon. With no consideration for any of the other touchpoints that influenced conversion which is clearly not correct and takes no real account of incrementality.
- Multi-touch attribution models
Multi-touch attribution models – such as Linear, Time Decay and U-shaped approaches – try to improve on single point models by distributing credit for conversion to a broader range of touchpoints across the customer journey.
However, the models are still relatively simplistic in nature and – like Single touch models – they also suffer from high levels of inaccuracy due to the poor quality of the underlying data. This is due to the limitations in the cookie/pixel-based approach being used with does a particularly poor job of ‘joining’ multiple user sessions across a range of devices – and generates data that is around 80% incorrect.
- Media Mix Modelling (MMM)
This approach applies statistical techniques to attempt to find incrementality in large data sets using mathematical techniques. Often relying on integrating panel data for offline activity.
In all cases, the object of MMM or Econometrics is to isolate a test case that is to be measured when it is applied in two or more variations. And the outcome is often expressed as a measure of incrementality to aspects of the media mix being used.
The challenge with this type of analysis is that the aggregated nature of the data greatly limits its effectiveness. And due to the complexity and volume of data required to create balanced tests and models that are statistically sound, media mix and econometrics studies done correctly are an expensive, ongoing proposition.
Related: A Comprehensive Guide to Attribution Models
It is frustrations with the limitations of these type of measurement solutions that has driven the development of new style attribution solutions that employ sophisticated Machine Learning and AI techniques.
These have the ability to unravel the complexity of multi-channel and multi-device journeys, rebuild poor quality clickstream data and break down the silos created by ‘walled garden’ type networks.
As well as take account of the contributions to conversion made by brand equity, offline media activity and existing organic behaviour to deliver true marketing incrementality.
How do these new solutions measure incrementality?
These new solutions deliver this enhanced level of analysis of incrementality by:
Rebuilding poor quality data
The reality is that poor quality data ‘in’ for analysis means poor quality data ‘out’ in attribution terms.
So, the priority here is to fully rebuild your raw clickstream marketing data from the ground up using session stitching Machine Learning processes. This completely rebuilds core marketing data from the ground up around a true picture of customer behaviour. Providing the basis for building a single view of each individual’s journey to conversion.
Unifying ‘walled garden’ and offline data
Using a probabilistic approach, it is then possible to join data from offline activity, store activity and even third party ‘walled garden’ digital data such as data from Facebook, Google Ads or YouTube.
For example, where predicted customer conversion behaviour is measured to have changed for a customer – and they have likely been exposed to an offline marketing event like a TV ad, or instore activity – then that activity is ‘stitched’ into their conversion journey.
Even at this stage, the improved data is significant in terms of the insight it can provide. For example, as a result of joining offline data to online, QueryClick were able to identify significant opportunities for spend reallocation in future TV and Radio media buying campaigns for a major life insurance customer. In fact, as the table below shows, 39% of TV spend was identified as ineffective and available for allocation to other channels for no negative revenue impact.
Attribute – and quantify true incrementality
While the customer engagement paths generated by the ‘Rebuild’ and ‘Unify’ phases are greatly improved data, they still represent a vast and complex relationship which must ultimately be scored to give a single attributed revenue outcome for each engagement point. Which considers the incrementality or otherwise of a piece of content or media on one particular, prospective customer journey.
Which is a true measure of incrementality for the media concerned – and more or less an achievement of the marketing holy grail we set out earlier in the blog!
True incrementality unlocks the key to revenue and ROI optimisation
So, what is the practical value in digging deep enough to identify true incrementality in your marketing activity?
Consider the example below, which is for a bricks and mortar and digital retailer which shows data for Facebook spend.
It shows the difference in the attribution analysis provided by Facebook and our own Machine Learning and AI driven attribution platform.
There are a couple of things that jump out of the analysis as follows:
- Facebook is over-reporting revenue by around 250% – due to a number of factors including the fact that their reporting is based on data generated from cookies which we know are notoriously poor in accuracy terms. In our own platform, AI and Machine Learning are effectively replacing cookies and doing a much better job of identifying individuals on their conversion path – boosting attribution accuracy from somewhere around 20% to 95%
- Spend re-allocation delivers 14 times more ROI – also by identifying the true incrementality in each campaign it was possible to identify that re-allocating £2k of spend could deliver 14 times more ROI.
This shows the true value of incrementality by providing proactive and predictive insight that is actionable to drive increased revenues and ROI.
And it is a principle that you can use to optimise your entire marketing approach at a marketing mix, channel, campaign and ad creative level.
Interested to learn more? Then download a copy of our Complete Guide to Marketing Attribution below.
The Complete Guide to Marketing Attribution
Own your marketing data & simplify your tech stack.
Have you read?
If there’s one thing SEO experts are used to, it’s turbulence and uncertainty. With Google’s AI Overviews arriving in the UK, a US judge ruling that Google’s search monopoly is...
According to Marketing Tech News, a staggering 92% of Marketers said that integrating AI into their marketing is the number one trend for 2024, and this trend doesn’t show any...
I have worked in SEO for 12+ years and I’ve seen the landscape shift a dozen times over. But the rollout of generative search engines (GSEs) feels like the biggest...