Subscription growth notes from 2,400+ Shopify brands | PMM @LoopEngine
Joined August 2025
- Tweets76
- Following151
- Followers17
- Likes110
one Loop brand generated $1M+ from upsells to existing subscribers at $0 CAC.
no new ads. no new traffic. just people already in the file who said yes to one more product.
~72% of that revenue came from the customer portal.
most brands treat the portal as a place to only manage subscriptions.
this brand treated it as a store.
existing subscribers are your highest-intent buyers -they already know the product. they already trust the brand. they already pay.
put a targeted upsell in front of them on the portal and you grow every order without touching CAC.
your next $1M isn’t hiding in a new ad. it’s sitting in the subscribers you already have.
RT @LoopEngine: More retries ≠ more recovered revenue.
We found that brands using personalized dunning campaigns recover ~30% more failed…
nailed it!
That’s the week-one onboarding. Put the same timeline in the customer portal, not only email.
Week 1: how to use it.
Week 4: what they should notice.
Week 12: the result they subscribed for.
Most first-order churn is an expectation problem.
working with a liquid dog supplement brand on retention
we didn’t start with winback flows, loyalty points or LTV cohorts
we started by rolling out a simple mnemonic across flows:
“Dropper. Dinner. Done.”
because the customer (in this case the dog) can’t get results from a product they forget to use
retention starts with consumption
your cheapest AOV bump is a subscriber who's already paid you for a month
show them the upgrade banner in Loop portal.
monthly → quarterly, 0 CAC, no new checkout
3 months of cash upfront
pump it straight back into your next meta ad campaign
Run the exclusive VIP BFCM offer to existing subs in the customer portal before the sitewide sale
How we run BFCM email marketing as an agency that has generated $300M+ from Emails 👇
Most brands treat BFCM like a 4-day sprint starting on Thanksgiving
In reality, it’s just the harvest phase of an 8-week operational machine
Here is the exact playbook we run:
Phase 1: The September Cleanup & Lock-in
- Hard deadline on the offer: Lock the hero offer before October 1st. Everything downstream (inventory, design, segmentation, copy) is blocked until this is signed off
- Ruthless list hygiene: Suppress all 90+ day unengaged contacts. You cannot repair a damaged sender reputation during peak week. Do it now so deliverability peaks in November
- Flow tune-up: Overhaul your welcome, abandon cart, and browse abandonment flows. Paid traffic will double in Q4, and your flows are the safety net that catches every dollar
Phase 2: October List Building & Modular Design
- Pack the list before CPMs spike: Swap standard 2% popups for interactive scratch-offs or mystery reveals (9-12% opt-in). Run traffic to a zero-discount "VIP Early Access" waitlist page
- Build modular templates: Never design fixed, single-image emails. Build modular blocks with live text so that when an inventory stockout or offer change happens mid-sale, it takes 20 minutes in Klaviyo instead of 2 days of redesign panic
- Kill the discount code: Build auto-apply checkout links into every button. Asking mobile shoppers to copy and paste coupon codes during flash sales destroys conversion
Phase 3: November 1-15 (The Runway)
- Zero discount fatigue: Stop running sitewide 20% off sales in early November. You train buyers to wait and burn out your list
- Send high-engagement plain text content: Founder stories, routine guides, and gift previews to train Gmail and Yahoo algorithms to place your domain straight in the Primary tab
Phase 4: Peak Week Execution (Nov 20 to Cyber Monday)
- VIP Early Access (48 hours pre-launch): Let your VIP tier and waitlist shop before the public. Generates massive immediate cash flow and clears stock before site crashes
- The Two-Send Rule: On peak days (Black Friday and Cyber Monday), run two sends per brand:
• Send 1 at 10 AM: Core announcement / stock alert
• Send 2 at 4 PM: Clean urgency push targeting non-openers and clicked-didn't-buy
(The 4 PM send routinely brings in 35-45% of total daily revenue with near-zero unsubscribes)
Phase 5: The Post-Cyber Week Cutoff (Dec 1-5)
- Extend Cyber Monday by 24-48 hours max for last-chance stragglers
- Clean break on December 5th: Kill all BFCM messaging immediately and transition 100% into holiday shipping deadlines and gift guides
If your strategy starts in November, you're competing on paid ads at $80+ CPMs
If you execute this roadmap, you harvest owned traffic on email for pure margin
Turns out you don’t need a crystal ball. Just ask Loop MCP 🔮
Subscription data shouldn't need a translator.
Loop MCP is live!
ask a question → get a straight answer. no dashboards, no cross-referencing, no waiting on a report.
running multiple stores? switch between them in the same chat.
want your own view of the business? build it with a prompt.
merchants have been on it for weeks: catching churn before it snowballs, spotting which cohorts are slipping fastest, making the call in minutes, not days.
one of our most requested features. also one of our favorites.
and we're just getting started. lots more coming to Loop MCP.
your subscription data. just ask 👇
one brand generated $210k revenue from subscribers it had already saved in the cancel funnel.
no ads. no new traffic. just people who clicked cancel, got saved, and ordered again.
most brands stop at save rate.
this one sits at 24% - on paper that’s $148k saved MRR.
save rate tells you who didn’t leave today. it does its job.
the metric that completes the picture: additional revenue generated.
revenue from orders placed after the save.
you know the cancel funnel is working when saved subscribers buy again.
Track the BFCM offer the same way you track the hook.
BFCM ads that win on CPA and lose on day-90 retention are just buying a cheaper churn cohort.
If the creative is “new sub deal,” exclude current subscribers from seeing it and give them an early-access gift instead of a cheaper rebill.
the best save last month wasn't a discount
Subscriber clicks cancel → selects "it didn't work for me".
Instead of an offer, they get one screen: the three reasons it usually underperforms, in plain language.
The last line of that screen is 50% off the companion product that fixes reason three.
The people who picked that reason in last 30 days, ~10% of them stayed.
You can't discount your way out of a product that didn't work but you can explain your way out of it.
Agrima Singh retweeted
One brand added $100k/mo in payment recovery.
From smarter retries.
Standard dunning logic : treats every failed payment the same. One schedule. One email. One retry cadence.
The problem: an expired card and an insufficient funds decline are not the same failure. They need different treatments.
The new logic:
- Expired card: Straight card update request. No urgency games. These subscribers intend to stay
- Insufficient funds (retry 1-5): Standard schedule. Most resolve within a week
- Insufficient funds (retry 5+): Apply a ramped auto-discount - 5% off per retry - to lower the payment threshold until the subscriber can clear it
Result on the brand above: $100k/mo additional recovery from the discount-ramping mechanic alone
The dunning schedule is one lever. Failure-reason-based logic is another.
Agrima Singh retweeted
Loop's at @beanstalkHQ from Sep 14-16 in Brooklyn.
Three days with the brands actually disrupting subscription commerce.
@piyush_jn & @sakshi3 on the ground the whole time.
Come talk payment recovery, churn, and customer portals that actually work for you.
Register here: loopwork.co/events/beanstalk…
Agrima Singh retweeted
New in Loop: instant winback offers 🔁
sub cancels → cancellation goes through → banner + countdown drops on their portal → they tap → offer opens → accept and the sub is live again
same cancel session. not a winback email 3 days later that nobody opens.
the good part: only people who actually cancelled, see it. so you can run a way steeper offer here than you'd ever put in your cancel-flow. nothing leaks to subs who were staying anyway.
set the validity window, cool-off settings, your incentive, your banner. ~1 min to go live.
Agrima Singh retweeted
Loop just shipped Mystery Rewards as a cancel save and it's a smart move.
Instead of the predictable "stay and get 30% off," the offer is randomized. Subscribers can't game the cancel screen for a coupon. That alone changes the economics of save offers.
The brands that will get the most out of this are the ones pairing it with upstream work. Frequency matching usage, product landing before renewal, acquisition bringing in people who actually want the subscription.
A stronger cancel screen plus a stronger front end. That's the combo.
@LoopEngine
_______________________
|
| the best subscription |
| brands don't retain |
| customers. |
|
| they build habits |
|____________________
\ (•ᴗ•) /
| |
Agrima Singh retweeted
We analyzed 441K cancel attempts to find out what separates 22%+ save rates from the rest.
Then turned it into a 3-min quiz.
8 scenarios. A score out of 200. your retention archetype revealed.
At the end:
→ see where you stand
→ find your blind spots
→ steal the full playbook - the cancel flow framework Loop brands use
Take the quiz. Reply with your score.
We're keeping track 👀
🔗 loopwork.co/interactive-quiz
Agrima Singh retweeted
A 12-month VIP and an order-2 subscriber eyeing the skip button shouldn't get the same reward.
But most retention plays hand everyone the same discount.
Mystery Rewards works by segment. Each subscriber sees a different surprise when they open the portal:
VIPs → full-size product, early access
At-risk → 20% off or a free sample
Order 2–5 → curated sample pack
They don't know what they're getting until they click.
That anticipation is what pulls attention off the cancel button.
And it's claim-based, not auto-applied:
No portal visit? No cost.
Portal visit but don't claim? No cost.
You only spend when someone actually engages. No blanket discounts eating your margin - just precision retention.
Agrima Singh retweeted
Introducing Mystery Rewards 🎁
Here’s what most brands don’t realize –
70% of your subscriber portal visits end in a negative action. Cancel, skip, pause, delay.
The portal was built for convenience. But it’s become the place where subscribers go to disengage.
Mystery Rewards flips that.
A subscriber opens the portal. Before they reach the skip button, a surprise is waiting.
A mystery gift.
That’s the "aha!" moment. Curiosity takes over. The cancel button fades into the background.
Personalized by segment. Claim-based offers so margins stay protected. No blanket discounts.
Brands using Mystery Rewards see ~50% reduction in negative portal actions.
Know more → loopwork.co/downloads/myster…