Marketing mix modelling: measuring what really drives sales
Marketing mix modelling (MMM) is a statistical method that estimates how much each marketing channel contributes to sales, by analysing historical data on spend and results alongside other factors like seasonality and price. It shows which channels drive revenue and how to allocate budget for the best return.
What marketing mix modelling (MMM) is, how it measures which channels drive sales, how it differs from attribution, and when businesses should use it.
TL;DR: Quick Answer
Basic South African brochure sites: R8,000-R20,000. Custom business websites with SEO and copywriting: R20,000-R50,000. E-commerce: R40,000-R150,000+. The five cost drivers that create the biggest price variation are: scope and number of pages, custom vs template design, professional copywriting, integrations (payment gateways, booking systems, CRM), and on-page SEO included at build stage. Always add 15-25% for hosting, maintenance and content updates in year one.
Key takeaways
- Very cheap quotes (under R5,000) almost always exclude copywriting, SEO, custom design and post-launch support
- Professional copywriting can represent 20-35% of a total website project cost, and is worth it for search visibility
- On-page SEO built into the website at launch costs a fraction of what it costs to retrofit after the site is live
- Hosting, SSL, domain and maintenance add R3,000-R10,000 per year on top of build cost
- E-commerce adds significant cost due to payment gateway integrations, product data, security requirements and checkout UX
- Timeline and client responsiveness directly affect cost: slow feedback rounds extend agency hours
What does marketing mix modelling do?
Marketing mix modelling uses statistics to untangle which of your marketing efforts actually drive sales. By analysing historical data, how much you spent on each channel and what sales followed, alongside factors like seasonality, pricing, and promotions, it estimates each channel's contribution to revenue.
The output is a clearer picture of return: which channels pull their weight, which underperform, and how shifting budget between them might change results. It answers the perennial marketing question, where is my money actually working, at the level of the whole marketing mix rather than individual clicks.
How does MMM differ from attribution?
MMM and attribution both try to credit marketing with results, but they work differently. Attribution tracks individual customer journeys, which ad or click led to a conversion, using digital tracking. MMM works at an aggregate level, modelling overall spend against overall sales.
| Aspect | Marketing mix modelling | Attribution |
|---|---|---|
| Level | Aggregate, whole mix | Individual journeys |
| Data | Historical spend and sales | Click and conversion tracking |
| Channels | Includes offline | Mainly digital |
| Privacy | Not reliant on tracking cookies | Affected by tracking limits |
The two are complementary. See our guide to multi-touch attribution for the journey-based view.
Why is MMM becoming more relevant?
Marketing mix modelling is having a revival because click-based attribution has become harder. Privacy changes, cookie restrictions, and cross-device journeys have weakened the digital tracking that attribution relies on, leaving gaps in the picture.
MMM does not depend on tracking individuals, so it sidesteps these problems. By working from aggregate spend and sales data, it can measure channels that attribution misses, including offline ones, and remains robust as privacy rules tighten. This is why methods once reserved for large advertisers are becoming relevant to a wider range of businesses.
What are the limits of MMM?
Marketing mix modelling is powerful but not magic. It needs a good amount of historical data to be reliable, so very new or very small marketing programmes may not have enough to model well. It also estimates contribution rather than proving it, since it infers patterns from data.
It is less granular than attribution, telling you a channel works without pinpointing the exact ad or keyword. And it requires statistical expertise to do properly; done badly, it produces confident but misleading conclusions. For these reasons, MMM suits businesses with meaningful, sustained spend across several channels rather than those just starting out.
When should a business use MMM?
MMM makes most sense for businesses spending substantially across multiple channels, including offline, with enough history to model. If you run TV or radio alongside digital, or simply spend across many digital channels and want to know the true contribution of each, MMM can guide smarter budget allocation.
For smaller businesses on one or two digital channels, simpler measurement, conversion tracking and cost per lead, is usually enough and more practical. As a business grows and its marketing mix broadens, MMM becomes a valuable tool for allocating budget across the whole picture rather than channel by channel.
See our guides to performance marketing and multi-touch attribution.
Frequently asked questions
What is marketing mix modelling?
Marketing mix modelling (MMM) is a statistical method that estimates how much each marketing channel contributes to sales, analysing historical spend and results alongside factors like seasonality and price. It shows which channels drive revenue and guides budget allocation.
How does MMM differ from attribution?
Attribution tracks individual customer journeys using digital tracking, mainly for online channels. MMM works at an aggregate level, modelling overall spend against overall sales, can include offline channels, and does not rely on tracking cookies. The two are complementary.
Why is marketing mix modelling becoming more relevant?
Click-based attribution has weakened due to privacy changes, cookie restrictions, and cross-device journeys. MMM does not depend on tracking individuals, so it sidesteps these problems and can measure channels attribution misses, making once-niche methods relevant more widely.
What are the limits of marketing mix modelling?
It needs substantial historical data to be reliable, estimates rather than proves contribution, is less granular than attribution, and requires statistical expertise. Done badly it produces confident but misleading conclusions, so it suits businesses with sustained multi-channel spend.
When should a business use MMM?
When spending substantially across multiple channels, including offline, with enough history to model. For smaller businesses on one or two digital channels, simpler measurement like conversion tracking and cost per lead is usually enough and more practical until the mix broadens.