Picture a Monday meeting. On the screen is a dashboard, refreshed overnight, with last week’s return for every channel. TV is down. Paid search is up. Who in the room can act on that? The TV airtime was booked weeks ago. The search bids are already being adjusted, hour by hour, by Google’s own algorithm. The year’s budget was agreed back in January. The number has arrived, and nobody has a decision to make with it.
Daily and weekly Marketing Mix Modelling (MMM) is easy to buy now. The technology behind it can be excellent. My concern isn’t with how it’s built. It’s that it serves a need almost no organisation actually has.
The easy cases
Some organisations don’t need MMM at all. A business with a small budget and just a few activities – everything online, some spend on Meta and a pay-per-click plan on Google – is usually well served by attribution, sense-checked with platform lift tests. Any good adviser will screen those out. The harder case is an organisation that is exactly right for MMM, and is being offered it at a pace it doesn’t need.
Three layers of decisions
I find it useful to think of marketing decisions as a pyramid with three layers. Strategic sits at the top – the annual reset, where budget is set across brands, categories and major channels for the year ahead. Tactical sits below it – the quarterly update, where the marketing team and its agencies refine that plan as the year plays out. Operational sits at the bottom – the day-to-day calls on bids, audiences and creative. Each layer makes its decisions at its own pace.
The bottom layer is taken
The ad platforms optimise bids, audiences and creative continuously, using signals no Marketing Mix Model will ever see. They’re built for that job, and they do it automatically. A model that refreshes every day won’t do it better. At this layer, MMM has nothing to add.
The top layers move slowly, by design
Airtime is booked in advance. Agency plans, contracts and production all take time to change. Decisions at the top of the pyramid are meant to hold, so that the plan has a chance to work. This is where Marketing Mix Modelling earns its keep: setting the budget once a year and refining it each quarter – or sooner, if something fundamental changes, such as a new channel, a pricing reset or a shock to the market.
The gap in between
So a daily or weekly MMM is serving the space in between: decisions too big for the platforms’ algorithms, but made more often than any planning cycle. How many of those does your organisation make? For almost every business, the honest answer is none. The number arrives every day, and there is no decision waiting for it.
That isn’t harmless. The data costs the same whether you run the model once a year or every day. What a daily model adds is the subscription and, above all, the demand on the team’s attention. Every meeting spent discussing a number nobody can act on is time not spent on the decisions that matter. And the weekly numbers mostly move with noise anyway – I explain why in Marketing Mix Modelling is not an operational tool.
The question to ask
So before you sign up for a daily or weekly MMM, ask one question: what would we do differently tomorrow because of today’s number? If the honest answer is nothing, you’d be paying for a need you don’t have. Run the model at the pace your decisions are made. For almost everyone, that’s once a year, with an update each quarter.
If you’d like a second opinion on how often your organisation really needs its model, drop a line to [email protected].