Brian Balfour vs Sean Ellis: Growth Systems or Growth Experiments?
Comparing Brian Balfour's growth loops and Four Fits with Sean Ellis's experiment-led growth hacking: how each frames growth, and when to think in systems rather than in tests.
Growth teams often split between those who run many fast experiments and those who design the system that produces growth. Sean Ellis, who coined growth hacking, represents the first. Brian Balfour, founder of Reforge and former head of growth at HubSpot, argues for the second. People compare them to decide whether to test more or model better.
What is the core difference?
Ellis treats growth as a discipline of experimentation. Form hypotheses, run tests quickly, learn, and repeat. The cadence of learning is the engine.
Balfour treats growth as a system whose parts must fit. If the product does not suit the channel, or the model does not suit the market, more experiments will not rescue it. He emphasizes loops, which feed themselves, over funnels, which need constant fuel.
One asks what to test next. The other asks whether the thing you are testing can work.
What does Ellis say?
Sean Ellis coined the term growth hacking in 2010. Hacking Growth (2017), with Morgan Brown, lays out a process built around a cross-functional team, a north star metric, and a high tempo of experiments. He promoted ICE scoring and the 40 percent very-disappointed test for product-market fit.
His method is aimed at finding and exploiting levers across acquisition, activation, retention, referral, and revenue. It is practical and drawn from startup experience, and it is widely used as a template for growth teams.
What does Balfour say?
Brian Balfour was head of growth at HubSpot and later founded Reforge, an education company for growth and product professionals. His essays and programs describe the Four Fits, the contrast between funnels and loops, and the idea that a growth model must be designed as a whole.
He argues that teams hit ceilings when they run tactics against a misaligned model, and that channels have a life cycle. His writing is a practitioner's synthesis and teaching, built from companies he worked with and advised, and Reforge sells paid programs on these ideas.
How do they compare?
| Dimension | Balfour | Ellis |
|---|---|---|
| Unit of thinking | The growth system and its fits | The individual experiment |
| Core ideas | Four Fits, growth loops, channel life cycles | North star metric, ICE scoring, experiment cadence |
| Typical question | Do our product, channel, and model fit? | Which test will teach us most next? |
| View of tactics | Useful only inside a sound system | A main source of learning and gains |
| Source | HubSpot growth work and Reforge | Early-stage startups and growth teams |
| Main risk | Over-modeling without testing | Endless tactics on a model that cannot work |
| Best known for | Reforge programs and essays on the Four Fits | Hacking Growth (2017, with Morgan Brown) |
When does each one fit?
Ellis's approach fits when you have a working core and want to improve its numbers: onboarding, activation, referral prompts, pricing pages. The tempo of tests produces steady gains.
Balfour's approach fits when growth has stalled despite many tests, or when you are choosing a channel or business model. The Four Fits help diagnose why effort is not paying off.
Where the product is not retaining users, both agree that the first job is the product. Without it, neither a system nor an experiment can create growth.
What does this look like in practice?
A software company sells a high-priced product to enterprises and has built a content and paid-social engine that brings thousands of small-business leads. Ellis-style experiments on the landing page might raise conversions by a few points.
A Balfour-style diagnosis asks whether the channel fits the product: small-business leads from social do not match an enterprise sales model. The fix may be a different channel, such as outbound or partnerships, or a different product tier. The experiments are then run on a model that can work.
"Describe our market, product, main acquisition channel, and revenue model: [describe]. Using Balfour's Four Fits, tell me which fit is weakest and what evidence supports that. Then, using Ellis's approach, propose five experiments inside the most promising fit, score each for impact, confidence, and ease, and name the one metric I should watch."
Why it works: the fits identify where effort will pay, and the scoring picks which test to run first.
Can you use both together?
Yes, and they work in sequence: use the system view to choose where to focus, and experiments to improve within it. The common error is to run endless tests on a bad fit, or to model forever without testing.
Both want evidence. A loop that looks good in a diagram still has to produce numbers.
"Sketch a growth loop for our product in which the output of one cycle becomes the input of the next, such as user content, referrals, or data. Identify the weakest step in the loop, estimate how much each cycle would grow, and propose two experiments to test that step quickly."
Why it works: a loop turns a vague hope for compounding into a testable chain of steps.
Where to go next
Brian Balfour and Sean Ellis both sit in the Marketing & Sales category. For the originals, read Hacking Growth by Ellis and Morgan Brown, and Balfour's essays and Reforge material. Reforge sells paid programs, so use the free essays to judge fit. To test a pricing structure, Pricing Decision is a $99 tool built for it.
Frequently asked questions
Does Balfour reject growth hacking?
He argues that tactics alone have limits. In his writing, Balfour contends that growth comes from a system whose pieces fit together, and that a team running isolated experiments can plateau when the underlying fit is wrong. He does not dismiss experimentation. He places it inside a larger model, which is a difference of emphasis from the experiment-first framing associated with Ellis.
What are the Four Fits?
Balfour's framework says a durable growth model needs four fits to align: market and product, product and channel, channel and model, and model and market. If one is out of line, effort on the others is wasted. For example, a channel that works for low-priced products may not suit a high-priced one. The framework appears in his essays and in Reforge programs.
What is a growth loop?
Balfour contrasts funnels, which are linear and require constant input, with loops in which the output of one cycle becomes the input of the next, such as content that attracts users who create more content. The idea is that loops compound. It is a way of modeling growth from his essays and teaching, not a controlled research finding. Loops matter because a funnel needs constant spending to refill it, while a loop can keep producing inputs for itself once it is running.
What is ICE scoring?
It is a method associated with Ellis for ranking growth experiments by impact, confidence, and ease, each scored on a simple scale. It helps a team pick which ideas to test first. Its weakness is that the scores are subjective, so it works best as a conversation aid with results reviewed after the tests. The scores are guesses, so record them and compare them to results, which gradually improves the team's judgment.
Which should an early-stage team use?
The documented work does not rank them. A team without a working model benefits from Balfour's questions about fit before running many tests. A team with a working model and a need to improve conversion benefits from Ellis's tempo of experiments. Many teams will use both at different moments. A short audit of the Four Fits is a cheap way to find out whether more testing is likely to pay.
Can AI help with growth modeling?
It can help map a funnel or loop, list assumptions, and draft experiments. It cannot see your data. Use it to structure the model and generate hypotheses, then validate with real numbers from your product. Give it your funnel numbers, the stages of your loop, and your assumptions, and ask it to identify the weakest link. Then test its read of the weakest link with a small experiment before you change anything large.
Written by Gareth Hoyle. Last updated 8 October 2026. Part of the authority.md guides library.
More guides.
Daron Acemoglu vs Ha-Joon Chang: Do Good Institutions Cause Growth, or Follow It?
Comparing Daron Acemoglu's institutions-first account of national prosperity in Why Nations Fail with Ha-Joon Chang's historical critique of the institutions literature: two views of development, with the main point of disagreement. General education.
Bill Ackman vs Carl Icahn: Two Styles of Activist Investing. Not Investment Advice
Comparing Bill Ackman's public, thesis-driven activism with Carl Icahn's decades of corporate pressure campaigns: how each approaches companies, boards, and the public. Not investment advice.
Adam Smith vs Ha-Joon Chang: The Founder of Market Economics and One of Its Historical Critics
Comparing Adam Smith's Wealth of Nations and Theory of Moral Sentiments with Ha-Joon Chang's history-based critique in Kicking Away the Ladder: what Smith actually argued, what Chang adds, and where they differ. General education.