Members go quiet before they cancel. First MagicAI Engage results show a message can bring them back.

Too often, a gym finds out a member has gone when the cancellation arrives. The signs are usually visible earlier: visits get less frequent, then stop.

Today the first results of MagicAI Engage are published, the AI that reaches out to members when their routine changes. Available as part of the Magicline platform under the Sport Alliance umbrella.

The first month matters most. In check-in data from 431,325 new Magicline members, one in five who visited in their first four weeks did not visit in the next four. By the end of the year, just under half (47.3%) were still visiting. These are members who stayed under contract all year.

At two larger operators, Engage messaged members who had stopped visiting or were visiting less. A randomly chosen group of similar members was deliberately left alone, so the two groups could be compared. Two weeks later, between 5 and 26 more members out of every 100 had come back in the messaged group, depending on the operator and the situation. At one operator, 40 of every 100 members who had stopped visiting and got a message came back within two weeks, against 26 of every 100 who did not. Across all ten comparisons at the two larger operators, over 7, 14 and 30 days, the messaged group came out ahead. Coming back is the first step. The study measured return visits, not renewals.

A second analysis shows why it matters: members who had gone a month without visiting, nine months into their contract, were less likely to still be members after their first contract year than members who kept coming in. At one operator, 48 of every 100 stayed against 77 of every 100. At another, 61 against 92.

“Members rarely leave in a single moment. Their routine changes first, and that is when a well-timed message can still make a difference. We wanted to know whether reaching out at that point actually brings members back. At the larger operators we studied, it did.” – Karl Foster, Head of AI

Engage watches every member’s attendance. When a routine changes, it sends the operator’s own message on the channel the operator chooses, such as WhatsApp, email, SMS or push notification. Operators can hold back a control group to see whether it worked. Because it is built natively into Magicline, it works from live member data. When a member replies, MagicAI Chat can pick up the conversation: Engage identifies, Chat acts. Product development is ongoing.

About the numbers: The test ran at three Magicline operators (28 studios). “Came back” means at least one visit in the following 14 days. The effect held after adjusting for members’ earlier attendance, studio and calendar week. The gap was wider at the operator with the smaller test group and narrower, but still clear, at the larger one. The third operator is a single studio, and its group was too small for a reliable result. The membership comparison covers 446 members at two operators. Those members had also visited less earlier and were more likely to have given notice, so it shows a link, not proof. The attendance data covers 431,325 members at 1,752 operators whose agreements permit pooled analysis.

Find more facts and insights on member reactivation in our research paper here.