5 Ways to Measure Marketing ROI

By Madison Taylor
August 6, 2021
an open wooden door on a grassy hill in a forest with a gold light in the doorway

Marketing has always been a relatively ephemeral field to track — you might put an ad in the paper and see a spike in sales, but how do you know how much of that spike is thanks to the newspaper ad and how much is coincidental?

The rise of marketing technology has helped to track user behavior, especially on the internet, but it’s also brought with it a new set of challenges. Between identifying a problem and making a purchase, a customer might click on a search ad, run across your page through organic search, engage with your social media profiles, receive an email, and download a whitepaper. How do you determine which one is responsible for making the sale?

As a result, measuring marketing ROI has always been tricky, to the extent that some marketers think that putting a hard number on their investments is a fruitless undertaking. We completely disagree. With modern tools and the right analytical approach, marketing ROI can absolutely be measured — and should be. Here’s how to get started.

Single Attribution

Single attribution, also known as first-touch or last-touch marketing, is the practice of attributing your conversions to either the first or the last point of contact with your company and brand. For example, you might host a webinar that generates several leads. One of them is nurtured by the sales team through several steps, culminating in a product demonstration that results in a sale. First-touch attribution would credit the webinar with the sale, while last-touch would credit the demonstration.

Single attribution is easy to implement and measure, and it can provide useful data if your funnel is relatively narrow, as is the case with many B2B companies that don’t use a wide variety of channels to attract new leads. On the other hand, it doesn’t account for the process of lead nurturing or multiple channels. It’s also difficult to account for lead quality — your numbers might be skewed by a particularly large sale.

Single Attribution with Revenue Projection

One problem with calculating ROI is that you’ll always have unresolved marketing efforts — if your marketing efforts haven’t paid off yet but you’ve already spent the money, it’s difficult to determine how much revenue they’ll bring in six months or a year from now.

One solution is revenue projection, which makes assumptions about lead conversions based on how similar leads have performed in the past. If you participated in a trade show a year ago and a third of your prospects became qualified leads, it’s reasonable to assume the same thing will happen this year, so you might extrapolate your prospect numbers from this year’s trade show to calculate ROI.

Revenue projection gives you a more complete picture of your lead pipeline and uses quality rather than just quantity, but it’s limited by other factors that might change your effectiveness. Past performance isn’t always a good indicator of future revenue, so your numbers will always be somewhat imprecise.

Multiple-Touch Attribution

A multiple-touch attribution approach operates under the assumption that multiple touches from multiple channels were required to close a certain deal and attempts to assign a value to each touch. A common approach is to assign the revenue equally between each significant touchpoint, but some marketers choose to weight the share of revenue based on recency, role, or program type.

A multiple-attribution approach incorporates more points along the pipeline, giving more credit to the nurturing process, which is especially useful for long revenue cycles with many touches. The downside is that there’s an inherently subjective nature to determining which touchpoints had the strongest influence, which can bias your data.

Test and Control Groups

This approach involves applying a particular approach to one randomly selected group of prospects or leads while not applying it to another randomly selected group. The difference in behavior should be accounted for almost entirely by the approach you’re testing.

For example, if you make 1,000 contacts at a trade show, you might send 500 of them a follow-up email with a downloadable whitepaper in it, while the other 500 don’t get an email at all. If more of the former group continue down the pipeline, you can calculate the value of that whitepaper.

There’s no limit to what you can test with this approach — revenue, leads, site traffic, profit, and so on — but it’s an expensive approach. You can’t test every element of your marketing strategy at once, since you need to isolate each variable to gain useful data from it, and each experiment takes time and money to run and examine the results.

Full Market Mix

A full market mix model involves collecting data about every channel in your marketing plan and running statistical tests on those channels to determine which is the most effective. It’s a complicated process that can’t realistically be conducted in real-time — the more likely scenario is that you’d run this method as a sort of audit of the past year’s worth of marketing spend to make informed decisions going forward.

A market mix model is the most accurate way to analyze your data, giving you insight into the effectiveness of each program and the impact of all internal and external factors, but it’s very time-consuming and expensive to conduct.

Choosing The Right Approach

The level of detail you pursue in your ROI measurements will depend on what you’re trying to accomplish and the scale of your operations. A single-touch approach can expose glaring problems quickly and easily — you might find that people whose first touch is a paid search ad almost never convert and immediately cut your search ad budget. If you’re looking to completely revamp a multi-million dollar marketing budget, however, you might need to take a deeper look with a series of test/control groups or a full market mix analysis.

It also doesn’t hurt to run multiple analyses on the same data to see if they agree with each other. If the simple results are giving you the same answers as the more complex ones, you’ll know that the simpler option is a close enough approximation for your purposes.

Whatever approach you take, the important thing is to simply start trying things and see if and how they work. The more data you gather, the more you’ll learn about how your customers think and the better you’ll be able to plan your budget.