The Latest from the ABM Experts
May 5, 2020
Measuring the Hard-to-Measure with Marketing Mix Modeling
Written by Sangram Vajre
Sometimes in life, it’s the hard-to-measure things that make the biggest impact.
Can you quantify your love for family and friends, your artistic sensibilities, your passions? And maybe you don’t need to. You just know they matter.
But when it comes to marketing, not measuring the hard-to-measure things can lead to some seriously unwise decisions.
In this #TakeoverTuesday episode, special host Steve Watt sits down with Chris Engman, CRO/CMO of Proof Analytics, to learn how marketing mix modeling can help you better understand these hard-to-measures and spend your marketing budget wisely.
Here’s what we’re unpacking today:
- How marketing mix modeling works
- The problem with multi-touch attribution
- How marketing mix modeling allows you to spend smarter
How marketing mix modeling works
Chris: What you’re trying to do with marketing mix modeling is to find the combination of factors that you spend the money on.
So, how much are you spending per marketing tactics and how are your volumetric KPIs looking on a weekly or monthly basis? Going back two, three, four years in time.
What you’re trying to do is you’re trying to use mathematical methods to use those variables to actually syntactically create your sales curve.
If you find a way to use the subset of all these variables — what you spend money on and also things like MQL, site traffic and all that — and you can find what mathematical formulas can recreate your sales curve, then you’ve found what is driving what over a certain time horizon and with how much volatility.
You start with the data — revenue profit — and a second group consisting of where you spend your money in marketing. Things like how much you spend per month on social media ads, sponsored posts, email marketing, events and so on.
Finally, you add in the third set of data: the KPI’s you are collecting. Things like site traffic, qualified leads, social media followers, conversion measurements.
Basically, you’re taking all of this and running multi-regression modeling based on it.
The problem with multi-touch attribution
Chris: I almost became a math scientist before I deviated into entrepreneurship — particularly marketing and sales.
I’ve been glancing at analytics anyway for the last 20 years, seeing nothing I really believe. I probably have too deep of a mathematical background to swallow lightweight models.
In the B2B space, a lot of companies are seeing multi-touch attribution as the next avenue.
I hate to be the party pooper here, but a lot of companies in the B2C space have gone deep down that and are turning back because the realize it’s missing the long-term effects.
It’s not catching brand new PR, missing offline activities and has trouble with multiple devices. It has too many flaws.
Marketing mix modeling reveals trends that let you spend smarter
Chris: I want to share something really pivotal I’ve seen. So many B2B companies are completely stuck in a marketing-qualified-leads belief system.
Marketing is way bigger than marketing qualified leads.
And there is also underinvestment in driving your funnel and increasing the deal size of existing accounts.
There’s an overspend on lead generation — and probably 90% of listeners have that issue without even knowing it. Top management is driving this because they need something quantifiable and it’s easy to quantify.
Looking at tactics, PR is almost always underfunded in the B2B space, while events are almost always overfunded — we quite often cut the event budget in half.
Finally, the sales force is almost always overfunded, so there are too many feet on the ground. For us, as we grew from $3 to 9 million in contracts in two years, we significantly cut our sales force and moved that money into marketing.
All of a sudden, marketing is driving the business and not sitting in the backseat, flipping brochures.