Why SCM analysts deny existence of real control
When I was in 5th grade, my friends and I desperately wanted to ride a roller coaster. All of us were a little short when measured against the signage.
We wanted to tear down the height measuring sign. This sign made us look bad. Years later it dawned to us that in fact the sign was there as a safety.
One can’t wish a away reality even if it makes us look bad.
SCM practitioners and Bayesians have one thing in common. They deny the existence of reality.
SCM analysts deny the existence of a ‘real control market’ while Bayesian would deny the existence of one true fixed parameter. Both hate the concept of accuracy.
Everything is uncertain and unpredictable for them and that’s what they want you to believe too. In fact you will rarely find the word ‘accuracy’ in relation to any SCM or Bayesian methods.
But why do they do this?
Because acceptance of real control or a true fixed parameter, makes them look bad and in a way denial of reality / ground truth gives them legitimacy.
SCM analysts would say “There is no such thing as real control market”. But this is a absolute cop out.
It is a tactic to lower the bar because SCM is an inefficient method. The very word ‘Synthetic’ connotes that there has to be something real and organic. If there is no ‘real market’ why even require an antonym ?
A control market is real if:
✅ It exists in the world (not made up),
✅ It is not exposed to the treatment,
✅ And it follows similar pre treatment dynamics.
By this definition, control markets absolutely exist.
The “synthetic unit” is not a real market. It’s a convex combination of fragments of other markets. The weights are chosen purely to minimize MSPE in the pre-period.
At no point does SCM ask:
– Do these markets share the same characteristics?
– Do consumers behave similarly?
– Are institutional, competitive or pricing realities aligned?
Reality is replaced by fit. People are replaced by weights. Markets are replaced by linear combinations.
SCM doesn’t discover a counterfactual. It manufactures one.
And the moment you question whether that synthetic construct corresponds to anything real, the legitimacy of the method starts to wobble.
So reality must be ignored 🤡.
If “no real control exists,” then DiD, RCTs, geo experiments and even SCM collapse as ideas, which is obviously not the case.
If SCM admits that its synthetic market isn’t real,
and if Bayesian admit that there is a fixed true parameter value, then both lose their strongest claim – methodological authority.
So instead, Both SCM and Bayesian hide behind unnecessarily made complex mathematics and superfluous words like ‘Robust, powerful’ etc.
Different methods but same escape hatch 😅.
If you want to implement causal experiments that work in reality, we are running the Marketing Measurement Marathon course. We still have 6 seats left for both Causality and Experimentation course. Link to register here: https://lnkd.in/gd7k7jfG)
