Forecast Chewy's Autoship Loop Factor Twelve Months Out

Marketing Academy · Field Work●Growth Marketing

CoreForecast· 45 minutes

Chewy

Objective: Given three plausible loop-factor scenarios for a subscription-driven paid loop, calculate the loop factor for each and forecast the twelve-month user-count difference between a leaking loop and a compounding one.

You're modeling growth scenarios for Chewy's Autoship-funded acquisition loop ahead of a budget review. Leadership wants to see, in real numbers, what happens if the loop factor sits at 0.8 versus 1.15.

Using the lesson's loop-factor formula (conversion x yield x velocity), build a 12-month compounding forecast under three scenarios and identify the single lever most worth fixing first.

Before you start

What you'll need

Free path (everything below is enough to finish)

Google Sheets

FreeBuild the loop-factor calculator and the 52-week compounding forecast

Free, formula-driven, easy to chart and share

Paid upgrades (optional, faster/deeper)

Free spreadsheet modeling is sufficient for a forecast; a paid product-analytics tool only matters once the loop is live and you need real cohort data instead of assumptions.

Amplitude(optional)

FreemiumTrack real cohort-by-cohort loop velocity once the loop ships, instead of modeled estimates

Cohort retention and behavioral-event tracking replace the spreadsheet's assumed rates with observed ones

No access? Google Sheets with manually logged weekly cohort counts

The process

2 steps

Step 01 of 02

Calculating the loop factor from conversion, yield, and velocity

The lesson defines loop factor as conversion rate x yield rate x loop velocity, multiplied together, with 1.0 as the break-even line between leaking and compounding.

Given Autoship conversion 60%, referral-yield 25%, and redemption-velocity 55%, what is the loop factor, and is this loop leaking or compounding?

Google Sheets— A sheet with one row per scenario and columns for each rate plus a computed loop-factor cell.

Procedure

Enter three scenario rows: Conservative (55% / 20% / 45%), Base (60% / 25% / 55%), Optimistic (65% / 30% / 65%)
Multiply the three percentages in each row to get the loop factor
Label each row leaking (<1.0), flat (=1.0), or compounding (>1.0)

Sample output

SCENARIO      CONVERSION  YIELD  VELOCITY  LOOP FACTOR  STATUS
Conservative  55%         20%    45%       0.0495       Leaking
Base          60%         25%    55%       0.0825       Leaking
Optimistic    65%         30%    65%       0.127        Leaking (all below 1.0 pre-scaling)

Healthy

The analyst multiplies all three rates correctly and correctly labels every result below 1.0 as leaking, resisting the urge to round up.

Unhealthy

Averaging the three percentages instead of multiplying them, which inflates the loop factor and hides a leaking loop.

What this means

Raw per-user rates on a referral loop are almost always well below 1.0 individually; the compounding case only appears once you model volume scaling across many concurrent cycles, not a single pass.

So what do I do about it?

Symptom	Action	Effort
A single-pass loop factor looks tiny and discouraging	Re-express the model as weekly cohorts stacking on top of each other, not one linear pass	30 min

YouYou can do this yourself, no engineering access required.

Step 02 of 02

Forecasting compounding growth from a loop factor above or below 1.0

The lesson's UpGrowth citation shows a loop starting at 20 users growing 10% week-over-week reaches thousands within twelve months without new spend, while a one-time campaign is dead after week one.

Starting from 20,000 Autoship subscribers, project week-52 user counts at 10% weekly compounding versus flat 0% growth. What's the twelve-month gap?

Google Sheets— A new sheet tab with a 52-row weekly compounding formula.

Procedure

Row 1: 20,000 users. Column A = week number, Column B = compounding formula =B(prev)*1.10
Add a flat-growth column that stays at 20,000 for all 52 weeks as the funnel-only baseline
Chart both lines and read off the week-52 values

Sample output

Week 1:  20,000 (compounding) vs 20,000 (flat)
Week 26: ~247,000 (compounding) vs 20,000 (flat)
Week 52: ~3,041,000 (compounding) vs 20,000 (flat)
Gap at week 52: ~3,021,000 users, entirely from the loop, zero extra spend assumed

Healthy

The forecast shows the gap widening every week, not linearly, and the analyst notes this is illustrative math, not a guarantee real-world decay won't slow it.

Unhealthy

Presenting the week-52 number to leadership as a committed forecast rather than a best-case illustration of what compounding could do if the loop factor holds.

What this means

The gap between a compounding loop and a flat funnel is not visible for months — it looks unremarkable through week 10, which is exactly why teams give up on loops too early.

So what do I do about it?

Symptom	Action	Effort
Leadership wants to cut the loop investment after 8 weeks of unremarkable numbers	Show the week-8 vs week-52 chart side by side to make the compounding curve's timing visible	30 min

YouYou can do this yourself, no engineering access required.

Final deliverable

A 52-week compounding forecast comparing a leaking-to-compounding loop factor against a flat funnel baseline, with the single weakest rate flagged as the priority fix.

See a reference example

Sample output

ThredUp resale-loop forecast (excerpt)

Base case: 15,000 sellers, loop factor 0.09 per pass, scaled to +8%/week across concurrent cohorts
Week 12: ~37,800 sellers
Week 52: ~712,000 sellers (modeled)
Weakest rate: yield (only 18% of sellers list a second batch) — flagged as the first experiment

Success criteria

You're done when you can:

Loop factor is calculated by multiplying, not averaging, the three rates
The 52-week forecast correctly shows exponential divergence from the flat baseline
One specific rate is identified as the priority lever, with a stated reason