There was an increase in revenue from at least two types of publications in…

Concept — A category counts as having increased in a given year only when its entry for that year is strictly greater than its entry in the immediately preceding year; an equal entry is not an increase. A condition of the form "at least two categories increased" is therefore tested year by year: for every year that has a predecessor row, count how many categories rose and compare that count against the threshold of two. The earliest row of a table has no predecessor row, so it can never be tested.

Application — The four years from 2020 to 2023 give three testable years.

1. The testable years are 2021 (against 2020), 2022 (against 2021) and 2023 (against 2022); 2020 is dropped because the table records nothing before it.

2. 2021 against 2020 — journals 45 → 47 (up), magazines 35 → 40 (up), books 70 → 80 (up). Three categories rose, which clears the threshold of two.

3. 2022 against 2021 — journals 47 → 40 (down), magazines 40 → 42 (up), books 80 → 87 (up). Two categories rose, which clears the threshold.

4. 2023 against 2022 — journals 40 → 40 (unchanged), magazines 42 → 54 (up), books 87 → 79 (down). One category rose, which falls short of the threshold.

5. The years that clear the threshold are 2021 and 2022, so the count is 2.

Cross-check — The three tests laid side by side:

Watch out — The journals entry for 2023 repeats the 2022 value of 40. Equal is not greater, so journals did not increase that year; treating that repeat as a rise would wrongly credit 2023 with two rising categories.

Result — Exactly 2 years — 2021 and 2022 — saw at least two of the three publication types earn more revenue than in the year before.