Launching New Channels With Incrementality Tests

As a marketing nerd, I find launching a new channel for a client exciting. Normally, by the time that first new campaign goes live there’s been plenty of research and preparation, so I can get antsy to see what results will start coming in.

Amidst the excitement it can be tempting to set a new channel live targeting as much of an ICP audience as possible, but this approach may not always be the best way to run a new test.

Launching a new channel in a handful of select markets can be an effective way to measure that channel’s true impact to performance. Especially if the new campaign isn’t expected to drive direct response actions (leads or purchases).

By definition, a new channel is unproven. Given that, I often recommend that my clients don’t spend a significant amount of budget on an unproven tactic until we can feel confident in the return we’re expecting. One of the best ways to evaluate the genuine impact of a new channel is through an incrementality test.

While it may sound technical, incrementality testing can actually be fairly straightforward. Another, potentially more familiar, name for this type of test is a geo lift test.

Basically, instead of launching a campaign targeting an entire country like the US, this test would only target a handful of states. For example, I might launch a campaign that is only targeting Florida, Illinois, and Georgia. While that campaign is gathering data over a few weeks, I’d compare performance against New York, Pennsylvania, and Ohio since this combination of states has a similar total population to the test campaign.

It can help to be sure the customer demographics are also similar between test and control regions.

To keep this example simple, if the campaign regions generated 15 leads over a 4 week period, and the control regions generated 12 leads, then my calculation would be as follows:

(15 leads from test regions - 12 leads from control regions) / 12 leads from control regions = 25% incremental impact from this campaign.

If the campaign spent $1,000 the true cost per lead wouldn’t be $66.67 ($1,000/15), but rather $333.33 ($1,000/3) since it’s likely 12 out of those 15 leads would have filled out a form anyway.

Granted, this example is an oversimplification, but highlights a simple process that can be followed when testing new channels. Running an incrementality test at launch can help determine whether or not this new channel type is worth additional investment.

Have questions, considerations, or critiques? I’d love to hear them! If you’re reading this via email, just hit respond. Otherwise, you can find me on LinkedIn.