Physics-Then-Loops Growth Modeling
Map the fixed physics of your business first, then find the one underused lever inside them to change.
- Difficulty
- Advanced
- Time to result
- ~weeks to results
- Steps
- 4
- Confidence
- 95%
A two-step model for deciding where to invest in growth. First, map the immovable 'physics' of your business — the market, product, monetization model, and channels. Second, examine your growth loops and funnels to find the single biggest constraint or most underused lever that fits within those physics, and change only that one thing at a time. The discipline is refusing to act like a wizard who can bend the whole universe at once.
Origin
Widjaja developed this as an intuitive process later formalized while building Reforge's 'Data for PMs' program; she credits Reforge with naming the implicit step-1/step-2 structure she was already using.
Core principles
- 01You are not a wizard — the physics of your market rarely change, so start from what already works
- 02Physics = market, product, model, channels; enumerate them explicitly before optimizing
- 03Change one small parameter at a time so you can attribute the result
- 04Your biggest growth unlock is often an existing, underused asset — not a brand-new feature
- 05Stay rooted in reality: fix the biggest current constraint rather than betting on many variables at once
How to run it
- 1
Map the physics of your business
List the fixed elements: the market and its supply/demand sides, the product that connects them, the monetization model, and the channels (paid, push, word-of-mouth, and any real-world channel). These are the constraints you must operate within.
Pro tip Include non-obvious real-world channels — for Gojek, drivers in branded jackets riding around the city were a primary word-of-mouth growth driver as it expanded to new cities.
Watch out Don't try to move the physics themselves; the universe isn't exploding, so treat market and model as roughly fixed.
- 2
Inventory your existing levers
Within those physics, identify the assets and levers you already have but may never have used — especially capacity that's underutilized, like a driver's captive attention during a ride.
Pro tip Ask what element of your model you've under-leveraged; the answer is often a growth channel hiding in plain sight.
- 3
Find the biggest constraint or best lever
Choose the single point where the largest constraint sits or where an existing lever offers the best unlock, rather than spreading bets across many new variables.
Watch out Changing four or five parameters at once destroys attribution — you won't know what worked.
- 4
Change one thing and fit it to the model
Make one small change, verify it fits the existing physics (no new product or model required), and measure the loop's response before touching anything else.
Pro tip A lever-only change with no change to the physics is the cheapest, highest-confidence growth move available.
In the wild
Analyzing Gojek's model, the team realized the driver's captive time with a rider was an underused lever. They built a small service that checked, on allocation, whether the customer had ever topped up GoPay; if not, it messaged the driver to collect cash, deposit it into the customer's wallet, and earn a bonus. No change to the product or model — just a new use of an existing lever.
→ The tactic drove roughly 60% of GoPay acquisition once released.
For an AB InBev D2C brand, cancellation reason #1 was 'I still have too much beer.' Rather than build new products or reactivation emails, they added a pause button — a temporary solution to a temporary problem — fixing the exact constraint where users dropped off within the model's existing physics.
→ The pause button alleviated a large share of hard-to-reacquire churn (mirroring Airbnb's snooze-listing win).
Common mistakes
Trying to change the physics instead of a lever
Founders act like wizards, betting on brand-new features and shifting many variables at once. Markets and models rarely move; changing multiple parameters simultaneously prevents attribution and usually fails.
Overlooking existing underused assets
Teams reach for novel features while an already-owned lever (like driver attention) sits idle. The biggest unlocks often come from using what you already have differently, not from building something new.
Is it for you?
Best for
Founders and growth leads deciding where to invest limited resources across a business with clear supply, product, model, and channels.
Not ideal for
Teams with insane product-market fit who are purely scrambling on core features, or research-stage products whose physics are still undefined.
From the transcript
“step one like you have constraints... you have to think about the physics of the current market the product the model and the channels that…”
“step two is when you think about loops and growth funnels and the quantitative inputs to each loop does that fit into these physics”
“we had underutilized the driver's capacity to drive our growth”
“it was huge it was like 60 of acquisition once we released that”
“we would always change like one small thing at a time and make sure that it fit into the model”
“you always have to start from the fact that like we are not wizards”
From the episode
How to scrappily hire for, measure, and unlock growth
Crystal Widjaja, Gojek and Kumu