Choose the cheapest channel that has the player's consent, move to a more expensive one only when the cheaper one has not worked, and keep calls for the players whose value pays for them. That is a cascade, and it is more useful than arguing about which channel is better in general. Which messenger or text channel works in your market depends on your players, and I cannot tell you that from here. I can describe the structure that works whatever your market is.
Which channels exist for casino retention?
In Sellrise the channels are SMS, Viber, WhatsApp, email, Telegram and calls. Push and in-app go through your platform. You can send through our channels, billed at cost, or through your own.
They differ in three ways that matter for retention: what each contact costs, whether you hold consent for it, and how personal it feels to the player. Email and push are cheap and low-touch. SMS and messengers cost more per contact and feel more direct. A call is the most expensive and the most personal.
What is a cheapest-first cascade?
You reach the player first by the cheapest channel that works for them, and escalate only if nothing happens within a sensible time. In Sellrise this appears as a cheapest-first channel cascade in the retention scenarios, with the live queue surfacing each save for a person to approve.
The logic is about money. If a push notification brings a player back, there is no reason to pay for an SMS. If a message does not, a more expensive step may be justified, but only if the player is worth it. Value decides how far up the ladder a player is allowed to go. For how value is estimated, see what predicted lifetime value means.
Who should get a call?
Players whose expected value covers the cost of the minutes. In practice that is the early VIP candidates and the high-value players who are cooling off. For them a call is both the most effective and the most affordable use of time. For a mass of low-value players it is not, and a team that calls everyone burns its day for little.
In the demo data, a VIP candidate in the first week is marked for a reload offer and a call, while a cooling-off regular gets cashback by Viber. These are examples from demo data, not client results. If you want to improve the quality of the calls themselves, see how to score every call in a casino call centre.
How does consent shape the plan?
It decides which channels you may use at all, player by player. Marketing consent is not a general permission. A player may have agreed to email and not to SMS. The check belongs at the moment of sending, not only at the time you build the list, because the state can change between the two.
Responsible gaming adds a stronger rule. Self-excluded players and those flagged must not receive promotional contact in any channel. I wrote about this in responsible gaming and casino retention.
Rules differ by country and by channel, and I am not a lawyer. Ask your compliance team which channels are open to you in each market, and treat that as a hard input to the plan.
How do you pick the time?
Use the player's own pattern, not a fixed delay. A chain that sends the same message after the same wait to everyone often arrives when the player has already left or deposited alone. In the daily list in Sellrise each row carries a time to contact, so a message goes when it fits the individual.
If you have no model, a simple approximation helps: look at the hours when a player usually deposits and aim for the hour before.
Frequency needs the same care as timing. A player who gets a message from every channel on the same day will not feel looked after. Agree a contact limit per player per week across all channels, and let the daily list respect it. If two scenarios want the same player, the more valuable action wins and the other waits.
How do you know a channel works?
Not by opens or clicks. Measure the outcome you care about, such as returns and average turnover per player, and compare with a group that was not contacted. A channel that gets many opens and no extra turnover is entertainment, not retention. The test design is in how to run a control group test.
Also compare channels against each other for the same type of player, since a channel can look strong only because it is used for your most valuable players.
Does anything send automatically?
Not in our model. Sellrise suggests, and your team decides and sends. Automatic scenarios within your rules are switched on only after the pilot, once the result is proven on your players. I prefer that order, because a system that has not been checked against your data should not be messaging your customers.
What is a simple channel plan to start with?
- List the channels you hold consent for, and the share of players covered by each.
- Rank them by cost per contact.
- Set a rule for who may reach each rung, based on value.
- Add a stop rule for responsible gaming status and consent, checked at send time.
- Run one campaign with a held-back group to confirm the cascade pays.
To see how the offer, time and channel are chosen per player, read the page on offers per player. You can also open the demo or start a pilot.