ChatGPT was the most common tracked subscription among Kahoot!+ subscribers
Share of 1,741 U.S. Kahoot!+ subscribers with each subscription at any point from 2021–August 2026
Share of Kahoot!+ subscribers
Source: YipitData U.S. e-receipt panel. Lifetime co-payment does not mean the subscriptions were held at the same time. Retention tests used separate time-ordered cohorts.
Chart data
| Series | Share of Kahoot!+ subscribers |
|---|---|
| ChatGPT | 35.4% |
| Duolingo | 20.3% |
| Grammarly | 10.2% |
| Quizlet | 9.1% |
| Chegg | 6.4% |
Did earlier ChatGPT adopters retain Kahoot!+ less often?
No lower retention appeared in either small-cohort comparison. In the August 2024–July 2025 base cohort, 27.9% of 86 earlier ChatGPT adopters retained Kahoot!+ in August 2025–July 2026, versus 22.2% of 158 non-adopters. In a separate 2023–2024 base cohort followed from January 2025–July 2026, 22.4% of 76 adopters retained, versus 17.7% of 238 non-adopters.
Was high ChatGPT overlap unique to Kahoot!+?
No. ChatGPT appeared among 37.4% of Quizlet subscribers, 37.0% of Coursera subscribers, 30.5% of Chegg subscribers, and 27.8% of Duolingo subscribers. The 35.4% Kahoot!+ rate sits within that broader observed range.
Does the retention comparison prove that ChatGPT helps Kahoot!+ retention?
No. The retention groups were small, and subscribers who pay for ChatGPT may also be more likely to retain subscriptions generally. The two comparisons show no evidence of lower Kahoot!+ retention among earlier ChatGPT adopters, not a causal benefit.
Assumptions & Methodologies
We analyzed YipitData’s U.S. e-receipt panel. Kahoot!+ subscribers were identified across Apple, Google, Amazon Appstore, Stripe, and Square receipts; physical name collisions were excluded. Overlap covered 1,741 subscribers from 2021–August 2026 and does not require simultaneous payment. Two separate retention comparisons began with a Kahoot!+ payment cohort, classified ChatGPT adoption before the follow-up window, and measured a later Kahoot!+ receipt. Retention groups ranged from 76–238 subscribers, and some billing channels are not observed, so treat the exact retention rates as estimates. The results do not establish complementarity, substitution, product use, or causation.