Clicks and conversions provide a convenient approach to estimating the effect of an ad. Brands get insights into how many consumers have interacted with the ad and eventually made a purchase influenced by it. However, these KPIs fail to capture crucial nuances.
Many brands unknowingly spend millions on marketing that doesn’t change consumer behaviour. A consumer might click on your ad, but would likely have purchased from you regardless. Incrementality testing allows you to tease out the true incremental effect of your spend so you can budget with confidence.
The problem all marketers face
Technology has failed to serve commerce brands and today, marketers are forced to rely on inadequate metrics for crucial decision making.
Classic metrics such as conversion rate and ROAS fail to distinguish between whether an ad caused the sale, or if the customer would have purchased regardless of having seen the ad.
Especially last-click and other types of attribution models risk over-crediting ad performance, leading marketers to become over-confident in their media efficiency.
This leads to budgets being wasted on ads that are merely capturing existing demand instead of converting prospects into new customers. Today, brands have limited opportunities to understand their media efficiency from this perspective.
The solution to the problems that many marketers face is Incrementality testing, which isolates the true effect of advertising. A market, typically a country, is split into two groups that historically have behaved similarly to one another.
By exposing one of the groups to an ad and not the other one, the true incremental effect of the investment can be understood by comparing the consumer behavior across the two.
Incrementality testing 101
Traditional attribution (including technology-forward solutions such as MTA and MMM) struggle to prove causality. These approaches, when set up correctly, can accurately assign credits across investments but fail to prove the causal impact.
A surprisingly large number of conversions would have happened anyways, and as a result relying on only these measurement approaches risks inflating reported ad effectiveness. Without testing for incrementality, brands risk scaling ineffective campaigns and ultimately wasting money on ads that fail to drive new revenue.
So, how are technically advanced marketers scaling this technology? In a nutshell, incrementality tests are designed to compare the behaviour across two or more similar groups of consumers. In the simplest case where two groups are used, one group receives ads in a specific channel, while the other one does not.
Apart from this, all other marketing between the groups is kept the same. This simple set up of carrying out one pre-defined action allows marketers to understand the causal uplift of said change by monitoring the behavior across groups for a duration of time. There’s an art to defining the experiment itself, but it deserves a separate post.
What makes Dema unique from other vendors in the space primarily comes down to the data that we have access to. Dema continuously tracks orders on a deep profit level across both new and existing customers.
This allows us to deliver hyper-granular test results, which other vendors are incapable of. We enable brands to follow up on their experiment across a wide range of KPIs including gross sales, profit and new customers to provide a holistic view of the incremental effects of marketing.
The continuous inflow of data, also allows brands to follow the experiment over time, avoiding being led astray by aggregate numbers. This is crucial, since we all know that timing is everything when it comes to marketing.
Shifting from Correlation to Causation
Marketers have a tough job and deserve better tools that allow them to focus on the right things.
Incrementality testing helps them step away from guesswork and instead make decisions backed by data that truly matters. Here’s three ways marketers can get started today.
- Define a hypothesis for what works and what could be working better. Depending on the budget-split between the two, begin by experimenting with the one with the largest spend.
- You don’t have to set up an advanced experiment design to get a good read on what works. You could for example try pausing ads in a specific channel for a number of weeks and get a read on the overall impact of sales.
- Since you’re likely already advertising on either Meta or Google, you can use their tools such as Conversion Lift studies to better understand the return you get from your marketing dollars.
Key Takeaways
Not all conversions are incremental, many would have happened anyway. MTA and MMM are both designed to assign credit to different marketing touchpoints; however, both fail to quantify the causal relationship between marketing and sales.
This is crucial for marketers to understand, in order to avoid wasted ad spend. Brands that continuously run these tests are able to understand where to cut spend and most importantly, they get the confidence and clarity to reallocate spend to real revenue drivers.