Double Down on the Loyal Remnant
Ignore the churned majority and over-serve the small cohort that stayed
- Difficulty
- Easy
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 82%
Zoom's first real distribution event was a Walt Mossberg review in the Wall Street Journal, accompanied by a video he recorded personally, which delivered fifty thousand users overnight. Most of them left within weeks. Eric Yuan's response was to treat the residue as the asset. Even a hundred genuinely loyal users, in his framing, is enough: identify the most loyal, double down on making them happy, and network effects do the recruiting because delighted users bring in a lot of new ones. The tactics were personal rather than programmatic — maintaining direct relationships, giving VIP accounts or small gifts, and personally emailing customers who cancelled a nine-dollar-ninety-nine subscription to ask what he could do differently. Peter Gassner ran a parallel practice at Veeva, deliberately refusing customer satisfaction surveys early because talking to early adopters gave him a feel that a survey would have hidden.
Origin
Eric Yuan on what Zoom did with the fifty thousand users delivered by a Wall Street Journal review, most of whom left within weeks, told on Acquired.
Core principles
- 01A spike of signups is not a business; the residue after the spike is.
- 02A hundred genuinely loyal users are enough to build network effects on.
- 03Delighted retained users recruit new users for free.
- 04Metrics let you hide from small numbers; talking to humans does not.
- 05Every cancellation is a free diagnostic if you ask for it personally.
How to run it
- 1
Let the spike settle
After a launch or press moment, wait several weeks and measure who is still there rather than celebrating the top-line signup number.
Watch out The headline number will be dramatically larger than the real one; plan for that emotionally.
- 2
Identify the most loyal, not the most numerous
Within the retained cohort, find the users who are genuinely enthusiastic. A hundred of them is a workable base.
Pro tip Look at frequency and unprompted advocacy rather than account size.
- 3
Over-serve them personally
Maintain direct personal relationships with those users. Give them upgraded accounts, small gifts, and real access to the founder.
Pro tip Keep the relationship personal — a founder replying directly is the product feature you cannot copy.
Watch out Delegating this to a lifecycle campaign destroys exactly the signal you are trying to collect.
- 4
Email every canceller yourself
When a customer cancels, write to them personally and ask why and what you could do differently, regardless of how small the subscription is.
Pro tip Yuan did this for cancellations of a nine-dollar-ninety-nine-a-month plan and kept the practice as the company grew.
- 5
Talk to humans instead of surveying
Skip customer satisfaction surveys while the numbers are small. Direct conversations give you the feel of what is going on; aggregate metrics let you hide behind them.
Watch out When it is small you can sort of hide behind metrics, and it does not work.
- 6
Let network effects compound
Very happy users bring in new users. Fund the happiness of the small cohort and let their referrals replace the acquisition spend you did not make.
Watch out This only compounds where the product itself is multiplayer or naturally shared.
In the wild
Weeks before launch, Zoom had no idea how to get its first customer. Walt Mossberg evaluated the service, wrote a positive Wall Street Journal article, and personally recorded a video, delivering fifty thousand users overnight. Most left within weeks. Eric Yuan's response was that even a hundred loyal users was good enough: identify the most loyal, double down on making them happy, and let network effects bring in new users. He maintained personal relationships with them, gave VIP accounts and small gifts, and personally emailed anyone who cancelled the nine-dollar-ninety-nine plan to ask what could be done differently.
→ A retained core of a hundred users became the referral engine for Zoom's early growth.
Peter Gassner deliberately did not run customer satisfaction surveys at Veeva in the early days. His reasoning was that talking directly to early adopter customers gave him a feel for what was actually happening, and that a survey would have taken that feel away. Eric Yuan agreed immediately, and the hosts summarised the shared mechanism: when the numbers are small you can hide behind metrics and it does not work, but if you talk to the human you figure out what is going on.
→ Both companies read early customer truth from conversation rather than instrumentation.
Common mistakes
Chasing the churned majority
Spending the next cycle trying to win back the thousands who left ignores the hundred who stayed and who are the only people telling you what the product is actually good for.
Instrumenting before you understand
Running satisfaction surveys while the user base is tiny replaces the founder's direct feel with an aggregate that hides the specifics you need.
Treating small revenue as unworthy of founder time
Dismissing a ten-dollar-a-month canceller as too small to email personally discards the cheapest and most honest churn diagnostic available.
Is it for you?
Best for
Early-stage self-serve products that get a burst of attention and lose most of it.
Not ideal for
Enterprise businesses with a handful of large accounts, where no meaningful cohort dynamics exist.
From the transcript
“double down to make sure they are happy if they are very happy guess what you know network effects they are going to bring a…”
“maintain a very good personal relationship with them either vip account or send them a small gift”
“we never had customer satisfaction surveys for viva in the beginning i always thought if i talk to those early adopter people i will know…”
From the episode
Capital-Efficient Growth (with Zoom CEO Eric Yuan & Veeva CEO Peter Gassner)
Peter Gassner