Blogs

The Only Thing Preventing Full Automation of MMM is Adstock – And That’s a Good Thing. In a recent meeting with a prospective client, we were asked – “So why is it that MMM can’t be fully automated ?” Fully automated MMM is a beautiful dream and every vendor and client has aspires for it. But there is one stubborn obstacle standing in the way – Adstock Adstock is not just a mere statistical transformation. It is a belief about how advertising behaves over time and it opens up many other pertinent questions like: ▪️How long does ads linger? ▪️Does performance media decay in days or weeks? ▪️Is the carryover

Why Calibrating MMM with Experiments can make you lose trust in both It goes without saying that Incrementality Testing (Causal Geo tests, Brand Lift tests, and event study) are very powerful. So are MMMs. At Aryma Labs, we use both. Whenever MMMs are not possible, Experimentation can be the answer. 2 years ago, a QSR brand doing MMM with us said – “we launched this special campaign which ran for 2 months. Can MMM tell whether it moved the needle of our sales?” When all your marketing/media spend variables span 2-3 years and you just have this short duration campaign, MMM is not the answer. We then looked at Economics

When to use which Causal Experimentation Product – A Playbook At Aryma Labs, we have a problem of plenty. We have a suite of Causal Experimentation products. However, some of our clients and prospective clients often ask us “Can you tell me when it would be ideal to use DiDective vs AdstockITSA?” To clarify this, we have created a intuitive playbook that illustrates when each tool comes in handy and what are its potential use cases. To augment the user experience further, we have built (and building) various AI agents that understand that context of experiments and provides real time insights and feedback. Also it serves as an excellent repository

It takes two to tango – You cannot build an accurate MMM model without the client’s inputs. Just early this week we onboarded a new client. The client asked “Ridhima, so how much support does your team need from my marketing team?” My answer : “Well.. a little bit 😊” MMM is often perceived as a purely statistical exercise : collect data, run models and generate insights. In reality, MMM sits at the intersection of data, business context and decision history. If you remove any one of these, accuracy suffers. Here’s why client input is so crucial. 📌 Data never tells the full story At Aryma Labs, we have an

What a very high correlation of your total media spends with sales tells about your brand? If your Total Sales are very highly correlated with Total Media Spends, your “Organic” is probably very low. And that’s not a problem. Yes, one can’t assert this statement always, correlation is not causation. MMM should be used to confirm. But in our practice, we have often seen that a very very high correlation often results in low base. This is typical in domains like D2C, Quick commerce (Blinkit, Zepto – In India), Beauty and personal care products and even in App installs and Gaming. 📌 The phenomenon : If your total sales goes

How Do you know your adstock values are accurate? During our recently concluded Marketing Measurement Marathon Course, a participant posed some excellent questions. “How do we know we have the right adstock parameter values? There are some thumb rules that carry over TV should be (0.3-0.7), Digital (0-0.3), Print /Radio/OOH (0.1-0.4). Why TV has a higher carry over than say Digital?” Let me try to answer the above 📌 How do we know we have the right adstock parameter values? Depending on the adstock type – Geometric or Weibull, you will have one or two parameters to estimate. We know we have the right adstock parameter values when the MMM

Geometric Adstock Simplified in 6 Steps In our recently concluded Marketing Measurement Marathon course, a participant asked us if there was any easy to remember code implementation of adstock. We hence created an intuitive step by step illustration of the Geometric Adstock Implementation. Geometric Adstock is the ‘hello world’ into the world of adstocks. The main advantage of geometric adstock is its simplicity and helps build the intuition. It is easy to understand and implement because it only requires estimating one parameter – the decay rate. Mind you that Geometric Adstock is not the prevalently used in real practice (at least not by us) as its assumption of constant decay

Industry Benchmarks in MMM should not be a North Star In a recent exploration call, a prospective client asked us “What does the average ROAS and MROAS look like for media channels in our industry? Can we aspire to reach that average or exceed it?” I have a bit of a contrarian opinion on this. Industry benchmarks of ROAS and MROAS in an industry should not be a north star. Why? Because they average out the behavior of all brands in a category. But here is the paradox, every brand at the same time is trying to outperform the category !! So by definition, the “average” is not where a

MMM Updates: Why they often happen based on Budget, and not when they are actually needed !! One uncomfortable truth in Marketing Mix Modeling (MMM) is this: MMM model updates are often driven by project budget, not by statistical or marketing necessity. In many engagements, monthly or quarterly or even half yearly “updates” exist because the pricing model demands it, not because the model genuinely needs a rebuild. But we all know that MMM models don’t decay on invoice cycles. They decay based on data dynamics and marketing reality. 📌 Updates vs Total Rebuild Adding one more month of data and re-running the same specification is not the same as

Causal Experiments Don’t Give a Holistic Picture of Incrementality Couple of days ago, I came across a interesting post in which the following claims were made. 1) In case of MMM, Contribution ≠ Incrementality 2) MMM and Causal Experiments should hone in on ‘one truth’ 3) Experiments provide a narrow but stronger causal identification. I wouldn’t have written this post, but the OP blocked me and deleted my comments for politely disagreeing and pointing out errors in their LLM generated post and LLM generated rebuttals. So let me provide my POV on the above. 📌 Contribution ≠ Incrementality MMM is implicitly causal and the biggest incrementality test one can do.