Data study · Wikipedia readership in the AI answer era

Weekday Work-Hour Wikipedia Reading Fell 8.0% Since 2022, Nearly Twice the 4.2% Drop in Weekday Evenings

KGWritten by Kriszta GrenyoPublished October 20263 min read

English Wikipedia views during US business hours on weekdays fell 8.0% between 2022 and the past 12 months. Weekday evening views fell 4.2%. On desktop, the work-hour drop reached 10.0%.

Analysis of 166.6 billion English Wikipedia page views by hour of week, 2022 vs October 2025 to September 2026 · Wikimedia pageviews via BigQuery, October 2026

Key findings

  1. 1

    English Wikipedia views during weekday work hours (9am to 5pm US Eastern) fell 8.0%, from 23.65 billion in 2022 to 21.76 billion in the past 12 months.

  2. 2

    Weekday evening views (5pm to midnight Eastern) fell 4.2%, from 18.09 billion to 17.34 billion.

  3. 3

    On desktop, weekday work-hour views fell 10.0% while weekday evening views fell 5.7%.

  4. 4

    Weekend views fell 9.1%, so the work-hour drop is larger than the evening drop but similar to the weekend drop.

  5. 5

    The single hour with the largest drop was 10am Eastern on weekdays, down 10.0% across desktop and mobile web; the smallest was 7pm, down 2.2%.

Summary

Picture a marketing manager at a 40-person company who used to keep Wikipedia open in a tab to check terms before a client call. Now she asks the AI assistant built into her work tools. Multiply that by millions of office workers and you would expect daytime reading to fall faster than evening reading. The data says it did.

We grouped English Wikipedia page views by hour of the week and compared 2022 with the latest 12 months. Weekday views between 9am and 5pm US Eastern time fell 8.0%. Weekday evening views fell 4.2%, roughly half as much.

The gap is wider on desktop, where most office research happens: work-hour desktop views fell 10.0% against 5.7% in the evening. Weekend reading also fell 9.1%, so the decline is not limited to the office. The evening is the part of the week that held up best.

What we measured

Wikimedia publishes page views by hour in UTC. We summed English desktop and mobile web views for each of the 168 hours of the week in every month, then shifted them back 5 hours to approximate US Eastern time. English Wikipedia is read across the world, so these hours are a proxy for the US day, not an exact measure of US readers.

We compared calendar year 2022 with October 2025 to September 2026 for four time blocks: weekday work hours, weekday evenings, weekday overnight and weekends. Hourly totals come from a separate rollup that is not limited to articles, so they add up to slightly more than our article-level totals in other studies.

Work hours lost almost twice as much reading as evenings

-8.0%change in weekday 9am to 5pm Eastern views, 2022 vs past 12 months

Each block of the week lost readers, but not evenly.

Change in English Wikipedia views by time block, 2022 vs Oct 2025 to Sep 2026
2022 (billions)Past 12 months (billions)
Weekday 9am to 5pm
23.7
21.8
Weekday 5pm to midnight
18.1
17.3
Weekday midnight to 9am
19.4
17.7
Weekend
25.4
23.1

% change

Time block2022 (billions)Past 12 months (billions)% change
Weekday 9am to 5pm23.721.8-8
Weekday 5pm to midnight18.117.3-4.2
Weekday midnight to 9am19.417.7-8.6
Weekend25.423.1-9.1

Source: aisearchoptimization.net analysis of 166.6 billion English Wikipedia page views by hour of week, 2022 vs October 2025 to September 2026 · Wikimedia pageviews via BigQuery, October 2026

Weekday evenings stand out as the most stable part of the week. Work hours, overnight and weekends all fell by roughly 8% to 9%.

On desktop, the work-hour drop reached 10%

-10.0%change in weekday work-hour desktop views

Splitting by site shows desktop carrying more of the work-hour loss.

Change by site and time block, 2022 vs past 12 months
Time blockDesktopMobile web
Weekday 9am to 5pm-10.0%-6.5%
Weekday 5pm to midnight-5.7%-3.4%
Weekday midnight to 9am-8.6%-8.6%
Weekend-10.0%-8.8%

Source: aisearchoptimization.net analysis of 166.6 billion English Wikipedia page views by hour of week, 2022 vs October 2025 to September 2026 · Wikimedia pageviews via BigQuery, October 2026

Desktop work-hour views went from 9.81 billion to 8.82 billion. Mobile evening views barely moved, from 11.91 billion to 11.51 billion.

Mid-morning took the biggest hit

-10.0%change in weekday views in the 10am Eastern hour

Looking hour by hour on weekdays, the drop peaks in the late morning and is smallest after dinner.

Change in weekday views by US Eastern hour, desktop plus mobile web
-10.9-8.52-6.1-3.68-1.266am9am11am1pm3pm5pm7pm9pm10pm

% change

Hour% change
6am-8.3
8am-7.9
9am-8.9
10am-10
11am-9.8
12pm-9.2
1pm-8.3
2pm-6.6
3pm-6
4pm-5.4
5pm-4.8
6pm-3.8
7pm-2.2
8pm-2.7
9pm-4.1
10pm-5.1

Source: aisearchoptimization.net analysis of 166.6 billion English Wikipedia page views by hour of week, 2022 vs October 2025 to September 2026 · Wikimedia pageviews via BigQuery, October 2026

The 10am hour lost 10.0% of its views and 7pm lost 2.2%. Late morning is a busy time for office look-ups, and a time when many workers may have an AI assistant open.

What this means

If you sell to businesses, your buyers do their research during the workday. That window includes the hour where reference reading has dropped the most, which fits the idea of AI tools at work taking over quick fact-checks and comparisons. Content written for a busy buyer at 10am now has to win a mention inside an AI answer, not only a click.

Look at your own traffic by hour. If weekday daytime visits to your guides are falling faster than evening visits, that is a sign buyers are getting answers elsewhere during work. Our free AI visibility audit shows what those answers say about your business, and if you want help acting on it, that is what we do as an AI search optimization agency.

Limits

Wikimedia pageview counts include some automated traffic that its filters do not catch, and Wikimedia changed how it detects bots in 2025, which lowers counted views from that point on. Some of the change in any period may reflect better bot filtering rather than fewer human readers.

We count views of titles that had at least 5 views in a given month on a given site, which drops a very small tail of rarely read pages.

Hours use a fixed UTC minus 5 offset, so daylight saving time shifts each block by one hour for part of the year.

English Wikipedia readers are spread across many time zones, including Europe, India and Asia Pacific, so US Eastern hours are only approximate for US readers.

The weekend drop (9.1%) is similar to the work-hour drop, so the decline is not only an office effect.

Embed this research

Add the charts from this study to your page. The embed links back here.

<iframe src="https://aisearchoptimization.net/embed/data-studies/work-hour-wikipedia-reading-decline" title="Work-Hour Wikipedia Reading Down 8.0%, Evenings Down 4.2%" width="100%" height="720" style="border:0" loading="lazy"></iframe>

Cite this study

APA

Grenyo, K. (2026). Work-Hour Wikipedia Reading Down 8.0%, Evenings Down 4.2%. AI Search Optimization. https://aisearchoptimization.net/resources/data-studies/work-hour-wikipedia-reading-decline

Plain link

Work-Hour Wikipedia Reading Down 8.0%, Evenings Down 4.2% - AI Search Optimization - https://aisearchoptimization.net/resources/data-studies/work-hour-wikipedia-reading-decline

Download the data (CSV)

Frequently asked questions

Has Wikipedia reading during work hours declined?

Yes. English Wikipedia views on weekdays from 9am to 5pm US Eastern fell 8.0%, from 23.65 billion in 2022 to 21.76 billion in the past 12 months.

Did evening Wikipedia reading drop too?

Less. Weekday evening views from 5pm to midnight Eastern fell 4.2%, from 18.09 billion to 17.34 billion.

Which hour lost the most Wikipedia readers?

The 10am Eastern hour on weekdays, which fell 10.0% across desktop and mobile web. The 7pm hour fell only 2.2%.

Is the drop larger on desktop during work hours?

Yes. Desktop work-hour views fell 10.0%, from 9.81 billion to 8.82 billion. Mobile web work-hour views fell 6.5%.

See where AI search leaves you out

Run the free AI visibility audit. If you want the fixes done for you, the plan is a flat $299/month.