The too-quick answer to this question goes something like this: “Of course. We know how much we spend on marketing and can measure the effects it has on growth, and from those figures we can see what return the business receives for its investment.”
Thinking about the issue for more than a few seconds, though, reveals that proving marketing ROI is not quite that simple. Other data points must be considered such as ‘how many sales would a company lose if it didn’t have any marketing?’ and ‘how much did we grow because our market grew?’
In an upcoming Ask Me Anything webinar on the 11th of October 2018, marketing industry leaders including Jennifer Villalobos of NTUC Income and Damien Cummings, former marketing head of numerous household brands, will discuss these questions and many other factors involved in proving marketing ROI. Those interested in the topic can sign up and post your own questions here.
For those in search of a ‘quick fix’, though, below are a few ways that marketers can discuss marketing ROI which depend on who is asking for proof.
1) Who’s Asking: A colleague
Imagine, you are trying to increase the number of downloads of your amazing, new whitepaper without upping the ad budget. You and a colleague are trying to redesign the landing page to improve conversions.
You prefer a simple approach: A site with limited text, a single email field, and a big ‘download’ button in the middle. Your colleague, however, is convinced that an explainer video at the top and an integrated social media data capture widget would work better.
Both approaches sound reasonable, but how can you prove which one will generate greater returns to the business?
One widely-accepted way of comparing two different tactical approaches is to run them side by side using an A/B test. You can configure the website so that half of the web traffic is sent to the first design (the ‘A’) and half is sent to the second (the ‘B’).
After a few days, the results are compared, and the proven winner is the approach which had more downloads. Simple, right?
What to watch out for
While A/B tests are popular and numerous professional tools exist to help marketers run them, they are based on some relatively complicated statistical analysis. Because of this, there are several important rules to follow:
- Make sure the test randomly redirects traffic on the same day, same time. Any variation in visitors can skew the results.
- Use a significance calculator. In order to determine whether your A/B test really tested a difference you need to check it against one of the many ‘significances’ checking tools on the web (such as VWO’s calculator).
- Don’t ‘peek’. Set a total number of visits beforehand and stick to the plan.
- Avoid any more than two varieties. Some tools allow for A/B/C/… testing. Additional variations increase the probability of a false positive and should be avoided.
2) Who’s asking: Your boss
It’s annual review time and while an A/B test may get that pesky colleague off of your back (or you off of theirs), your boss may not be so easily convinced. It’s great that you got more downloads for one whitepaper, but how will they know whether your work has made a difference over many months and many campaigns?
Well, to prove your personal ROI you will need to have some agreed metric or metrics (though one is ideal) which can be used as a key performance indicator (KPI). It could be increasing web traffic through content marketing, click through rates with some snazzy ads, or the average revenue per visit with some clever merchandising.
Then, with a bit of data wrangling, you can show that, on average, the KPI was higher.
What to watch out for
Besides the very tricky question referred to above (whether the KPI would have changed anyway), you also have to worry about variance. Variance is the rate at which a metric changes naturally and it applies to averages as well as your KPI.
So, if you want to prove that your marketing skills truly ‘moved the needle’, your results will have to be greater than the previous average plus the variance.
Here’s a post which explains how to calculate variance yourself using R, or if you have Excel 2016, you can use its forecast sheet to see whether it was your work or pure chance which boosted your KPI.
3) Who’s asking: The CFO
It’s the annual budget meeting and the chief financial officer (CFO) is looking at two numbers on the income statement, ‘Gross Profit‘ (the return) and ‘Sales, General and Admin‘ (the investment, often dominated by marketing costs). Using these numbers, the CFO has a pretty good idea of the overall return on the investments that the company has made in its various overhead or fixed costs.
Luckily, the return is greater than the investment so at least there is some ROI, but unfortunately for you, other departments have already claimed that it was their hard work which resulted in the returns, not marketing’s.
What can you do?
Proving marketing ROI to CFOs is tricky, but fortunately there are ways chief marketers (and people trying to make chief marketers happy) can prepare for it.
1) Review IPA case studies
The Institute of Practitioners in Advertising holds annual Effectiveness Awards which require contestants to submit detailed campaign and business results to qualify. While few, if any, will match your marketing strategies, the cases are presented in a logical, business leader friendly style and provide a great template for a marketing ROI argument. The Economist’s Raising Eyebrows and Subscriptions is particularly good.
2) Get familiar with Binet and Field’s landmark reports
Chief marketers should download Binet and Field’s reports, The Long and the Short of It and Media in Focus. Both speak at length about how long-term brand building is preferable to short-term activation for having very large business effects on key business metrics. Learn from their deep thinking and clear writing on the topic of marketing ROI.
Econsultancy’s Maximising the ROI of Digital Marketing report
3) Keep meticulous records of marketing activity
As Byron Sharp notes in his provocative text, Marketing: Theory, Evidence, Practice, collecting marketing activity records ‘provides marketing analysts with useful raw material to investigate how your markets and marketing works’.
Sounds obvious, and yet, the text continues, ‘managers often fail to systemically record their own marketing activity’ and that it is ‘very difficult, or impossible, to generate [marketing activity] data on the past if it wasn’t collected at the time.’
Now, there is no fixed template for proving ROI once you have your marketing activity data, but without it, you will struggle to build a case with the same persuasive powers as the ones you ready about in the IPA case studies.
So, sadly, there is no ‘quick fix’ for marketing heads trying to prove ROI but reading about the state of the art on the practice and keeping track of your, and your competition’s, marketing activities should help you prepare for the inevitable reckoning day with the CFO.
For more insights on this topic, join Econsultancy’s Proving ROI: Ask Me Anything webinar on 11th October 2018.
Econsultancy is also offering a Proving Digital ROI workshop on 16th November 2018 in Singapore. Click here for more information and to book your spot.
And you always need to remember these 5 things about testing…
(1) The “novelty effect”, aka “any damn fool marketing idea will probably work for a bit”. It’s an opposite effect to “ad blindness” and means anything new is likely to work better than what you’re doing now. This is why I dislike “control groups” in marketing.
(2) The answer to “are your results significant?” is always, “significant for what?” If you’re sending the best of 2 similar newsletters, then an 80% chance of sending the right one is ample, because you don’t *really* care about the result and there are advantages in reducing the size of the test groups. But if you are deciding between two expensive paid campaigns, then you want much higher certainty. Here’s what Jeff Bezos says,
If you want the rest of the five, I wrote a blog here…