How to build a growth process (growth loops) instead of one-off marketing campaigns

AI advisor. Alikhan

Brief description

You run ads, get a spike in leads, and then everything goes quiet. To keep going, you have to pay again.


Details

Level: for beginners

Language: Russian

Duration: 1:46


Lesson transcript

You run ads, get a spike in leads, and then everything goes quiet. To keep going, you have to pay again. Most teams know this situation, and it has a name: a campaign. The alternative is called a growth loop.

The idea was popularized by Brian Balfour, founder of Reforge and former VP of Growth at HubSpot, in his article “Growth Loops are the New Funnels.” His point is simple: the fastest-growing products are better described not as a funnel but as a system of loops. A funnel is linear: a user enters at the top, some drop off along the way, and that's the end of it. To get a result again, you have to invest money again. A loop is closed: the output of one cycle becomes the input for the next.

Every loop has four parts. Input: a new user. Action: what they do in the product. Output: what this action creates, such as content, an invitation or revenue. And reinvestment: the output feeds back in and brings in the next users. If the output doesn't return to the input on its own, it is not a loop but a campaign.

A classic example is Pinterest. A person signs up, saves pins and builds boards. Public boards are indexed by search engines. New people find this content through search and sign up themselves. Each user helps attract the next ones without a separate budget.

How to build this kind of process yourself. First, find an action that your best users already take and that leaves a trace outside the product: a post, a report, an invitation to a colleague, a link. Then remove friction: the fewer steps between the action and a new user appearing, the faster the loop. Next, measure two things: the share of users who complete each step and the length of the full cycle. The shorter the cycle, the faster growth compounds. Then find the weakest point, where the most people drop off, and improve exactly that.

An important condition: a loop only works with retention. If people leave after the first time, there will be no one left to invite others or publish, so the retention metrics we discussed earlier remain the foundation.

One more clarification: the funnel doesn't disappear. Conversion from visitor to customer is still a funnel, but now it is built into the loop. Paid advertising can also become part of a loop if revenue from customers pays back and funds the next ad purchase.

The bottom line: stop asking how much money you need for the next campaign. Ask how the results of your product bring in the next customers on their own.

Cheat sheet: growth loops instead of one-off campaigns

How a loop works, how it differs from a funnel, and how to measure and build it. Calculations use a hypothetical example.

1. The key idea in one sentence

A growth loop is a closed system in which the output of one cycle becomes the input for the next. If the output doesn't return to the input on its own, it is not a loop but a campaign.

2. What a loop consists of

PartWhat it isExample (Pinterest)
InputA new user or a unit of contentA person signs up
ActionWhat they do in the productSaves pins, builds boards
OutputWhat this action creates outside the productPublic boards show up in search results
ReinvestmentThe output brings new people to the inputNew people find the boards through search and sign up

3. Funnel vs. loop

FunnelLoop
ShapeLinear: input, losses, endCircular: the output returns to the input
GrowthMore money in, more results outThe faster the cycle, the faster the growth
Budget dependenceNeeds a constant new inflowPartly grows on its own
RolePart of the loop (visitor-to-customer conversion)The whole system

The funnel doesn't disappear: it stays inside the loop as the conversion stage.

4. Types of loops

  • Viral: a user invites others (invitations, links, referrals).
  • Content / search: users create content that attracts new people through search (as with Pinterest).
  • Sales: satisfied customers provide reviews and case studies that help close the next deals.
  • Paid (self-funding): revenue from customers pays back and funds the next ad purchase.

5. Loop metrics

  • Conversion at each step — what share moves on to the next one.
  • Cycle time — the median number of days from a user's input to the appearance of a new input.
  • Loop velocity — how many rounds the loop completes in a period: days in the period / cycle time.
  • Viral coefficient K (for a viral loop) = invitations per user × invitation-to-sign-up conversion.
  • Retention — without retention, the loop doesn't work: users who leave stop creating output.

6. Sample calculation

Each month, 1,000 new users come from ads. Each of them invites 5 people on average, and 10% of those invited sign up. The cycle takes 7 days.

Viral coefficient: K = 5 × 0.10 = 0.5.

Loop velocity over 28 days: 28 / 7 = 4 rounds.

How many people the loop will bring in from the initial 1,000:

RoundNew users
Initial1,000
1st1,000 × 0.5 = 500
2nd500 × 0.5 = 250
3rd250 × 0.5 = 125
4th125 × 0.5 = 62

Total over 4 rounds: 1,000 + 500 + 250 + 125 + 62 ≈ 1,937. If the loop keeps working, the limit is 1,000 / (1 − 0.5) = 2,000.

How to read the result: with K below 1, the loop doesn't grow on its own, but it amplifies paid acquisition: each paid user brings in another half a user for free. With K above 1, each cycle yields more than the previous one, and growth becomes self-sustaining.

What to improve: raising invitation conversion from 10% to 14% gives K = 0.7 and a limit of 1,000 / 0.3 ≈ 3,333. Cutting the cycle from 7 to 5 days increases the number of rounds per month from 4 to 5.6.

7. How to build a loop: step by step

  1. Find an action that your best users already take and that leaves a trace outside the product: a post, an invitation, a report, a link.
  2. Draw the loop: input, action, output, return.
  3. Remove friction between the action and a new user appearing.
  4. Measure the conversion of each step and the length of the full cycle.
  5. Find the weakest step, where the most people drop off, and improve it.
  6. Check retention: if users leave, the loop will stop.
  7. Add a second loop only once the first one works.

8. Common mistakes

  • Treating a campaign as a loop: the output doesn't return to the input.
  • Trying to invent a loop from scratch instead of strengthening an output that already brings in new people.
  • Improving all steps at once instead of a single bottleneck.
  • Ignoring retention.
  • Measuring only the number of new users, not the cycle time.

9. Checklist

  • I can name the input, action, output and return in my loop
  • I know the conversion at each step
  • I know the cycle time
  • I know the weakest link
  • Retention is high enough for the loop to work

Sources:

  1. Reforge (Brian Balfour) — Growth Loops are the New Funnels
  2. Amplitude — Growth Loops Explained
  3. Ortto — Growth loops: How to use them and real-life examples
  4. Searchlab — Growth loops vs funnels

Author
Alikhan
Alikhan AI advisor
Timecodes
0:00 One-off campaigns
0:12 Funnel vs loop
0:33 Four steps of a loop
0:43 The Pinterest example
0:59 Launching your own loop
1:12 When loops work

Please take 5 minutes to leave us a review

Оставьте обратную связь