How to build retention flows in Telegram and email that pay back
Telegram and email flows that pay back are segmented by behavior, triggered by actions, and measured against a holdout group.
Clean data and a way to measure incremental revenue
Retention flows only work with clean data and a real way to measure incremental revenue. Without both, every result is a guess dressed up as a report.
That starts with identity: a reliable way to pass subscriber ID from your funnel into your messaging platforms, email, Telegram bot, whichever you're using, via webhooks or API integrations set up at the point of purchase or signup. If you can't tie a message back to a user's first touch and last purchase, stop here and fix that before anything else. The rest of the minimum setup:
- User ID or unique identifier passed at every event: purchase, signup, trial start
- Event tracking on the key actions: first purchase, repeat purchase, churn, reactivation
- UTM or source tagging so you know which channel drove the subscriber
- Cost-per-acquisition (CPA) per channel, updated weekly at minimum
- A clear definition of 'active' versus 'churned' for your product, e.g. no login for 30 days
Segment your audience by behavior and value
Segment before you send anything - sending the same message to everyone causes retention flows to fail. The simplest split runs on recency and monetary value: three core buckets for email and Telegram, active high-value (purchased in the last 30 days, LTV above median), active low-value (recent but low spend), and churned (no purchase in 60+ days).
A SaaS product with a free trial adds a 'trial started but not converted' bucket; e-commerce splits 'one-time buyer' from 'repeat buyer'. Each bucket runs its own flow, with its own frequency and offer - a churned customer gets a re-engagement discount, a high-value customer gets VIP content or early access.
One EdTech course we ran Telegram retention for segmented by 'completed module 1' versus 'not started'. The 'not started' group got a 3-message sequence with study tips and a direct link to the first video, and open rates hit 45% against 22% on the generic broadcast. Reactivation cost per subscriber came to $0.30 versus $1.10.
Design the flows: sequence, timing, and triggers
Every flow needs one clear goal: bring the user back to a specific action. Map the journey backwards from there - a second purchase means the flow should start 3-5 days after the first one, while a churned user gets a reminder 7 days after their last session.
Trigger on behavior. A user adds something to cart and doesn't buy: send a Telegram reminder within 2 hours. A subscription expires: run an email sequence on day 1, 3, and 7 with reactivation offers. Keep Telegram sequences short regardless - 3-5 messages max, then stop unless the user engages again.
Timing matters by channel. Telegram is more intrusive, so stick to active hours, 10am-1pm and 6pm-9pm local time; email can go out any time but should skip weekends for B2B. Test the send times themselves - we've seen CPL for email reactivation vary by 40% between a morning send and an evening one.
Write copy and design creatives that match the channel
Telegram copy stays short, direct, personal: use the user's first name, short paragraphs, one clear CTA per message. 'Hey {name}, your trial ends in 3 days. Click here to keep access.' reads right; text and a link only.
Email has more room but still needs one focus. Subject lines personalized with name or product name lift open rates by 20-30% in our own tests. Plain text works better for re-engagement emails: it reads more personal and pulls higher reply rates. Transactional emails, receipts, shipping notices, can carry a subtle upsell link.
A/B test the CTA text on both channels: 'Claim your discount' against 'See your offer' can swing click-through rates by 15-25%. Test one variable at a time to make the result meaningful. Sender name matters too: on Telegram, a real person's name, like 'Anna from Support', pulls 35% more replies than a bot name.
Set up tracking and measure payback
Every message needs a trackable link: UTM parameters with source=telegram or email, medium=retention, campaign=reactivation_60d, and the user ID passed right in the link so conversions tie back to the specific message that drove them. A redirect server like Keitaro or Binom, where possible, counts the clicks and postbacks on top of that.
Payback per flow is revenue attributed to the flow minus the cost of sending it and the cost of any discount inside it. Cost of sending covers platform fees - email ESP cost per 1000 sends, Telegram bot hosting - plus creative production. Take a typical SaaS product at $50 a month: a re-engagement flow that generates 10 extra subscriptions at $50 each brings in $500 in revenue. A $50 sending cost leaves $450, and the flow still clears its own cost even at $200 to run.
Attribution windows run 7-day or 30-day; shorter windows overattribute to last touch, which is why we use a 14-day linear model for retention flows in e-commerce.
Optimize based on cost per action, beyond open rate
Cost per click and cost per conversion drive revenue. An email sequence with a 50% open rate and zero conversions is worthless; a Telegram flow at 10% open rate but 5% conversion is the one you keep funding.
Optimize by testing segments, send times, and offers against each other. A holdout group makes the result real: hold back 10% of users from the flow at random, then compare their behavior to the group that got it. The gap between them is the incremental revenue.
One mobile app test ran retention flow against no flow directly. The flow group posted 12% 30-day retention at a cost of $0.08 per user per month; the no-flow group's natural retention sat at 8%. That 4% lift was worth $3,200 in extra LTV per 10,000 users, against a flow cost of $800 - a payback of 4x.
Automate the lifecycle end to end
Automation scales beyond a handful of users. Automation tools cover it: Mailchimp, ActiveCampaign, or SendGrid for email; Manybot, Chatfuel, or a bot built on the Bot API for Telegram, with triggers set on user events: purchase confirmation, abandoned cart, subscription renewal, inactivity.
Map the full lifecycle: welcome flow for the first 7 days, engagement flow across days 8-30, at-risk flow at 30-60 days inactive, churn flow past 60+ days, each running 3-5 messages. The at-risk sequence might run day 30: 'Haven't seen you in a while'; day 37: 'Here's a tip to get started again'; day 45: 'We miss you, here's 20% off your next purchase'.
Telegram buttons beat links for quick replies - a 'Yes, I'm interested' button that triggers a discount code can double conversion rates against a plain link. Even the button text matters: 'Get my 20% off' against 'Claim offer' showed an 18% difference in clicks in our testing.
Scale across channels and geos
A flow that works in one channel can adapt with channel-specific adjustments: Telegram users expect immediacy, email users expect more detail. Localize the timing: send at 10am local time, and vary the offer by geo where price sensitivity differs (10% off in the US, 15% off in the EU).
Scale the segment size while keeping frequency steady. A flow that works for 1,000 users might work for 10,000, but test the jump gradually and watch for list fatigue - open rates below 10% on Telegram or 15% on email mean the list is saturated, and that segment's flow needs a 30-day pause.
Budget 20-30% of total marketing spend for retention once there's critical mass - on a $10k/month ad spend, that's $2k-$3k going to retention tools and content. The return should run 3-5x that within 60 days.
Six ways retention flows go wrong
- Sending too often means more than 1-2 messages a week on Telegram leads to 40%+ unsubscribe rates; on email, cap it at 2-4 per week unless the message is transactional.
- No segmentation, just blasting everyone the same message instead of splitting by behavior.
- Ignoring churn in favor of chasing new users only, when churn flows can recover 5-15% of lost users at a low cost (CPA $1-$3).
- No tracking means plain links with no UTM or postback, which means optimizing blind.
- The same copy on both channels, when Telegram needs short and urgent while email can carry a longer story.
- Never testing the offer itself: a 10% discount can underperform a free shipping code with the exact same audience.
Open rates, reactivation cost and how long results take
Results take 2-4 weeks to show meaningful data. Standard non-transactional numbers run 20-30% open rate and 3-5% click rate for email, 30-50% open rate and 5-15% click rate for Telegram, though both vary by vertical.
On cost: an email ESP for 10,000 subscribers runs $50-$200/month, Telegram bot hosting $10-$50/month, and creative production for 10 messages $200-$500 if outsourced or free in-house. A full retention program serving 50,000 users lands around $500-$2,000 a month total, depending on tools and content.
On return: properly segmented and tracked flows deliver 2-5x ROI within 90 days. E-commerce retention flows can contribute 15-25% of total revenue; for SaaS, 10-20% of monthly recurring revenue can trace back to automated retention messages, with churn reduction of 10-30% achievable on top of that.
FAQ
What's the best channel for retention: Telegram or email?
Both work, for different audiences. Telegram runs higher open rates, 30-50%, and feels more personal, but it can be intrusive; email sits lower at 20-30% but is expected and handles larger lists with less operational overhead. Running both, with different content for each, gets the best results.
How often should I send retention messages?
Telegram tops out at 2 a week unless the user has engaged recently; email is safe at 2-4 a week. Let the user set their own frequency in a preference center where possible - over-sending is what kills list health fastest.
How do I measure ROI of a retention flow?
Track the revenue from users who received the flow, subtract the cost of sending plus any discounts, and use a holdout group to measure the incremental lift on top of that. Attribute revenue on a 14-day window.
What's a good cost per reactivation?
In Tier 1 geos, cost per reactivation, a churned user making a purchase again, should run $1-$8: $1-$3 for email, $3-$8 for Telegram. Compare that to new customer acquisition cost; retention should come in at least 50% cheaper.
Can I use the same flow for different verticals?
Each vertical runs different purchase cycles and user expectations. Dating apps need gentle reminders; iGaming needs urgency. Tailor sequence length, offer, and tone to the product every time.
I can do this on your product
I consult on acquisition, funnels and retention - including hard verticals.
