Rolling Growth Forecast Loop
Model the path to growth, test its assumptions, and update it as evidence changes.
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 93%
Convert an ambitious growth target into a forecast that explains where each increment of growth is expected to come from. Allocate contributions across relevant mechanisms such as organic search, paid acquisition, partnerships, and virality, then state the assumptions behind each contribution. Compare actual results with the forecast frequently to determine whether the data supports those assumptions. When evidence changes, update the model, revise the hypotheses, and redirect resources rather than defending the original spreadsheet. Because today’s successful channel may not scale indefinitely, the loop repeatedly searches for additional scalable mechanisms while preserving a visible path between current activity and the long-term target.
Origin
Kieran Flanagan proposed a frequently updated forecast model while explaining how a new Loyalty Lion marketing leader could pursue repeated high growth. Extracted from Marketing Against The Grain.
Core principles
- 01Every ambitious goal needs a plausible path.
- 02A forecast is a collection of testable assumptions, not a fixed promise.
- 03Repeated growth requires finding mechanisms that continue to scale.
- 04The model should change when data invalidates a hypothesis.
How to run it
- 1
Define the target
Specify the growth outcome, time horizon, baseline, and measurement definition.
Pro tip Separate year-over-year growth from absolute volume targets.
Watch out An ambiguous target cannot support a useful model.
- 2
Construct the path
Allocate expected growth across channels, programs, conversion improvements, and other mechanisms.
Pro tip Show the arithmetic relationship between each component and the total target.
Watch out Do not fill unexplained gaps with a generic other category.
- 3
Record assumptions
Document traffic, conversion, cost, capacity, and timing assumptions for each modeled contribution.
Pro tip Mark the assumptions with the greatest uncertainty.
Watch out Hidden assumptions make forecast errors difficult to diagnose.
- 4
Compare forecast with reality
Review actual results on a regular cadence and isolate which assumptions are holding or failing.
Pro tip Evaluate leading indicators before final revenue arrives.
Watch out Do not wait until the end of the year to discover that the path was wrong.
- 5
Revise and reallocate
Update the model and move resources toward validated mechanisms while designing new tests for unresolved gaps.
Pro tip Preserve prior versions so the team can learn from forecast error.
Watch out Changing the model to conceal underperformance destroys its value.
In the wild
A marketing leader models growth from organic search, paid acquisition, partner distribution, conversion improvements, and customer referrals. Monthly evidence shows paid acquisition saturating earlier than expected while partnerships outperform. The leader updates the forecast and reallocates budget and staff toward partner programs while testing a replacement for the paid shortfall.
→ The plan remains connected to evidence instead of becoming an obsolete annual spreadsheet.
Common mistakes
Treating the model as a commitment
A forecast should expose and update assumptions, not pressure the team to preserve predictions that evidence has disproved.
Extrapolating current channels forever
Mechanisms that work today may saturate or weaken, so sustained growth requires continued search for scalable additions.
Is it for you?
Best for
Marketing leaders accountable for sustained multi-channel customer or demand growth.
Not ideal for
Very early products with no usable baseline data and no immediate way to test demand assumptions.
From the transcript
“to get anywhere, you have to have a path. You have to kind of know how you're going to get there, right?”
“When I would come in, I would probably really nail my forecast model, try to figure out where I expect that growth to come from.…”
“When you're trying to 10x numbers, you have to repeatedly find things that scale, right? And the things that are working today may not work…”
From the episode
The CMO Show. How to Find One. How to Be One.