A Guide to win-back sequence in Web3 Email Marketing

A win-back sequence is a set of messages sent to somebody because they stopped at the moment they stopped.

That is the whole idea. Ecommerce has run it as standard practice for over a decade. Almost nobody in this market runs one, and the reason for that is worth getting to. Start with what it's actually worth.

Omnisend published its 2025 email dataset: 20 billion campaign emails, 470 million automated sends, more than 27,000 brands [1]. Inside it, customer reactivation emails opened at 33.11%. Ordinary marketing campaigns opened at 30.41%. Now look at what happened after users opened. Reactivation converted at 0.54%. Campaigns converted at 0.08% [1]. That’s almost the same attention, but close to seven times the action.

The question you might be asking is, why?

Let’s picture the person on the other end, receiving the email; let’s call the person Ark. So Ark has bought from ShopQut every three weeks for about a year, and then Ark stopped. Ark did not complain. Ark did not unsubscribe. Ark just went quiet, but ShopQut kept sending Ark the same Tuesday newsletter, because as far as ShopQut is concerned, Ark is just one name in fifty thousand they have on their list. However, Ark doesn’t open it because there’s no more interest.

Six weeks later, a different email arrives in Ark's mail. It comes because ShopQut has identified that Ark stopped buying. The email copy is no smarter, and the offer is no bigger. That email carries one piece of knowledge the Tuesday newsletter does not: It’s smart, personal and driven by data.

According to the Omnisend report, sends built on these three pillars convert nearly seven times better.

The same is true for blockchain companies. A wallet (not Ark now, haha) deposits $5,000 into a vault. It stays five months. In month six, the deposits stop. In month eight, the wallet withdraws everything and never comes back.

Every step is visible, live, down to the transaction. At most blockchain companies we have studied, when the user stops, the engagement with the company ends. With no win-back sequence in place to make it work.

This is a significant gap in how much value blockchain companies can generate. If they can retain the users they have, here’s what I mean.

The cheapest user in your funnel is the one you already have

Every funnel you have built spends money at the top.

Ads, influencer marketing, quests, incentives, content, airdrop and points campaigns are all aimed at making a stranger care enough to try out your product or chain.

A win-back sequence works at the other end of the funnel, on somebody who already cared enough to use your product or chain. They onboarded. They read the docs. They trusted you with money. Then something changed; they stopped. In marketing, this is called churn. And win-back sequences are one of the most effective ways to win back churned users.

That is the benefit. Here is the system for how it works:

Threshold

The threshold is the line that says a user is churned and needs a win-back sequence. It has to be a number with a unit, agreed in writing before any user crosses it, because "inactive" means five different things to five people on the same team.

You do not pick that number out of the air. You take it from how your own users normally behave.

1. A coffee subscription sets their threshold line at 45 days, because a bag lasts about a month, so somebody who has not reordered in six weeks is buying coffee somewhere else.
2. A mattress shop that copies the same 45-day line is emailing people who are perfectly happy with the mattress they bought last month. See?

For a blockchain company, the unit changes but the thinking does not. Say you run a lending market where suppliers normally adjust a position every few weeks. Two lines you could write as thresholds could be for when:

1. Supplied balance has fallen more than 70% below that wallet's own 30-day average, or
2. An open position has not been touched in 30 days.

This is how the threshold works. If you set the threshold and 350 wallets are already sitting below it.  win-back sequence is triggered, and those 350 carry the rest of this example.

Trigger

The trigger is what fires when a user crosses that threshold, and the important word is when.

A campaign goes out on Tuesday to whoever was on the list on Monday. However, triggers are different; they watch the threshold all the time and fire for one user the moment that user crosses it, at whatever hour that is.

Based on this, you’d notice that the trigger is a response to the threshold, which is a measure of the user behaviour (when they do x, do y). This means Win-back sequences are actually behavioural in design, making them smart, personal and data-driven.

For the 350 users who triggered the threshold in the analogy used, they all enter the flow on day one, because they crossed before anyone was watching. After that, the trigger works one wallet at a time. Wallet 351 crosses at three in the morning on a Sunday, and gets its first message at three in the morning on a Sunday.

That is the difference between a trigger within a win-back sequence and a campaign. A campaign goes out on specified days, to whoever was on the list.

Sequence

The sequence is the arrangement of the response to the threshold being triggered. Let’s say the supplied balance has fallen more than 70% below that wallet's own 30-day average; 350 users who meet this threshold.

The trigger is activated, and all 350 users receive 1 email instantly to let them know and another email 24 hours later, including an In-app push to any of the users who log into the app within the first 48 hours. Everything that happens after the trigger is pulled is called the Sequence.

Usually there are three steps, each with a different job.

1. A knock on the door, with no offer. Its job is to find out who still opens anything from you.
2. The reason to come back, has to be a change rather than a plea. A rate that moved, a fee that dropped, a thing that was broken when they left and works now.
3. The goodbye. If nothing has landed by now, its job is to remove them from the list.

A fourth step is worth adding for the ones who did come back, carrying an upsell or a reward. Three is the floor rather than the rule, and the frequency of the sequence is yours to define.

Channels

Now channels are the mediums through which the messages in the sequences are delivered.

If you remember the previous analogy, where all 350 users receive 1 email instantly to let them know and another email 24 hours later, including an In-app push to any of the users who log into the app within the first 48 hours.

In this analogy, 2 channels were used. One thing to know is that each channel reaches a different state, which is the part most teams get wrong:

1. Email is the only channel that reaches somebody who has stopped visiting you, which is exactly who a win-back is for.
2. In-app push reaches the wallet that came back and then stalled, and it cannot reach anyone who never returns.
3. SMS and WhatsApp cost real money per message and interrupt harder, so they earn their place only in your highest-value segment.

You can, just like in the analogy, run all channels as one sequence with one rule: if the user acts on any channel, the rest stop. Without that rule, a user clicks the email and deposits on Tuesday, and your in-app message still asks her to come back on Thursday.

What it costs, against what churn costs

Now each of those messages costs you to send. In the Omnisend report, reactivation earns the least of any automation, $0.51 per email against a $3.41 average, and loses 0.68% of an already-lapsed audience [1].

Against the campaign it replaces at $0.155, it still makes 3.3 times as much per send [1].

When you look at the blockchain industry and how we treat retention, you’d instantly know that it could get very expensive and yet not be effective. In October 2024, Optimism sent 10.4 million OP to 54,700 addresses. An analysis on the governance forum compared wallets that just cleared the 50 OP line against those that just missed it: clearing it raised 30-day retention by 4.2 percentage points, and by day 60 that was 2.8 [3].

That’s four points, a third of it gone within a month, paid in tokens. Compared to the effectiveness of airdrop campaigns as a retention system to drive repeat users after testnet, a proper winback sequence could be more effective and cheaper. So the question is, why is almost nobody doing it?

Blockchain companies can watch the churn and cannot respond to it

When you go back and read through the four parts that make a win-back sequence work, try to ask yourself: what does a blockchain company need for each one? A win-back sequence has four parts:

1. A threshold: The point at which someone is considered inactive or at risk of leaving
2. A trigger: The moment they cross that threshold
3. A sequence: Usually three messages designed to bring them back
4. A Channel: Typically email and in-app messaging

For a blockchain company, defining the threshold is the easy part. Onchain behaviour is public, so a protocol can decide exactly what “at risk” means: a user’s deposit falls below a certain level, they stop borrowing, they move funds elsewhere, or they have not interacted with the product for a set period.

The hard part is acting on that signal immediately.

To do that for an individual wallet, a system needs to continuously monitor onchain activity across multiple networks and protocols. It needs to recognise that a withdrawal on Aave and then a swap on Uniswap from one wallet may all represent the same behaviour: a user reducing their relationship with your product.

Today, most teams handle this manually. They run a query in Dune or another analytics tool like Allium or Token Terminal, then export a CSV, match wallets to emails, and upload the list into an email platform. By then, the opportunity has often passed. The user showed signs of leaving on Wednesday, but did not receive an email about their action till tuesday.

Another challenge: wallets are not email addresses. To contact someone, a company needs to connect a wallet to a person, collect consent, and be able to remove that connection if the person asks. Many teams get stuck at this step, and that’s why they never launch a proper win-back programme.

Then there is channel coordination. A strong sequence should work across email and in-app push, from one workflow. Building that internally requires significant data and messaging infrastructure. When you think about the hurdle it will take to implement a standard lifecycle system for blockchain companies, you’d instantly understand why these companies can watch the churn and cannot respond to it. That is why we built OnchainSuite.

OnchainSuite creates one loop with six stages:

• Detect captures onchain activity and product events as they happen
• Resolve connects wallets, emails and external IDs into one contact record
• Segment creates cohorts using rules or SQL across onchain and messaging data
• Reach sends email and wallet-addressed in-app push from the same workflow
• Automate triggers flows based on wallet actions, including mints, swaps, approvals and liquidations
• Measure tracks delivery, opens and clicks across both channels

The mapping to a win-back sequence is simple:

For example, a lending protocol can set a rule once: when a user’s deposited balance drops below a defined threshold, start a three-step win-back sequence through email and in-app push.

From that point, our system responds to each wallet individually when its behaviour changes, at any time of day, until the user returns or the workflow is paused.

Why onchain win-back could perform better

Onchain win-back should have the potential to outperform traditional ecommerce win-back.

An ecommerce business often works with a limited signal: the date of someone’s last purchase. A protocol can see much more than that; they can see:

• The size of a user’s position
• Whether that position is increasing or shrinking
• How quickly funds are leaving
• Whether the wallet has started supplying, borrowing or trading somewhere else

That makes the second message in a win-back sequence much more relevant. Rather than simply saying, “We miss you,” a protocol can explain why returning may be useful based on the user’s actual behaviour.

The limitation is reach. Onchain audiences are pseudonymous, email coverage is still low for many protocols, and consent must be collected before messaging. If a protocol can only contact a small percentage of wallets that leave, even an excellent sequence will have limited overall impact. That is why two questions matter more than opinions:

1. Does a wallet-balance-based trigger reactivate more users than a purchase-date-based trigger?
2. How long is the quiet period between a wallet’s last meaningful action and their actual exit?

That quiet window is the opportunity. A wallet that becomes inactive in month eight is not so different from someone who stops buying coffee from the same shop. Both go quiet. Both may be reachable for a limited period.

The difference is that, onchain, the behaviour is visible in real time and the protocol can act on it. If you’d love to try out proper retention and lifecycle flows, try out https://onchainsuite.com/.

References

1. Omnisend, "Email Marketing Benchmarks: Open Rates, Clicks, and Conversions," 2025 dataset. https://www.omnisend.com/blog/email-marketing-benchmarks/ 
2. Klaviyo, "2026 Email Marketing Benchmarks by Industry." https://www.klaviyo.com/products/email-marketing/benchmarks 
3. Google, "Email sender guidelines," Gmail bulk sender requirements. https://support.google.com/a/answer/81126 
4. "Did OP Airdrop 5 Increase User Retention Rates? A Regression Discontinuity Analysis," Optimism Collective governance forum. https://gov.optimism.io/t/did-op-airdrop-5-increase-user-retention-rates-a-regression-discontinuity-analysis/9610 
5. OnchainSuite documentation. https://docs.onchainsuite.com/