Research:Readers' research 2025
This page documents a research project in progress.
Information may be incomplete and change as the project progresses.
Please contact the project lead before formally citing or reusing results from this page.
This research will investigate the behaviors and motivations of Wikipedia readers in 2025 in order to identify potential product and community interventions that will enable Wikipedia to serve as a source of useful information for readers in the future.
The Wikimedia Foundation notes several ongoing trends related to readership, including a gradual decline in overall visitors (e.g. 3.8% year over year decline in user pageviews from April 2024 to April 2025), and major changes in the ways people use search, social media, and AI tools. A new generation increasingly uses visual, short-form platforms to pursue information, and they particularly prefer information aggregated by people and personalities they trust. Search platforms are likewise heavily investing in AI-driven experiences. People are increasingly using ChatGPT and other LLMs — and while AI for information search still lags behind traditional web search, reliance on it is growing. In general, trust in information online is declining, and shared consensus may be fragmenting.
WMF product teams working on reader growth and retention feel they lack a deep understanding of readers and their needs, to ensure they are working on the most effective product improvements, and to determine whether there are aspects they haven’t considered, or new initiatives or specific features they should prioritize. Generally, the group also feels they lack clear frameworks and shared vocabulary to describe what user needs readers have when visiting Wikipedia. This shared vocabulary would help facilitate faster and easier internal discussions and alignment.
This research specifically aims to support the Wikimedia Foundation objective for Fiscal Year 2025-26 WE3, which anticipates a future state in which “readers from multiple generations engage, and stay engaged, with Wikipedia, leading to measurable increases in retention and donation activity”.
Methods
editResearch Goals
- Revisit and flesh out the taxonomy of reader use cases to create a usable, memorable framework for the product team that names and describes the use cases that Wikipedia is used for, by readers
- Identify and understand what’s most important for readers for each use case (what does a successful session look like, in their eyes?) as well as what are the biggest current pain points to solve for
- Understand when and why Wikipedia is used, vs. other apps/sites used, for these use cases – across a broad swath of people that ranges from people who read Wikipedia regularly, to those who do so less often or not at all
Specific Research Questions
Use cases
Use case demographics
- Which demographic differences are associated with engaging in or avoiding specific Wikipedia use cases?
Use case behaviors
- What reader behaviors are associated with specific use cases and readers’ perceptions of their success or failure (e.g., session length, number of pages visited, reader goals, etc.)?
Use case modality
- For each use case, are there modality differences in which use cases are more or less common on each platform? (desktop vs. mobile web vs. mobile apps)
When and why Wikipedia vs. other information sources
Wikipedia vs LLMs
- To what extent are AI./LLM tools such as ChatGPT used for identified Wikipedia use cases?
Wikipedia vs other knowledge platforms
- What strengths and weaknesses do users associate with each of these platforms?
- Why do they prefer certain apps/sites as their go-to, for certain use cases? (specifically what attributes of that app/site do they like?)
Wikipedia in the information diet
- Where does Wikipedia fall in participants’ information ‘diet’? What role(s) does it play? Why?
- What do people see as the biggest strengths or differentiators of Wikipedia, compared to other information sources?
Who doesn’t use Wikipedia?
- More specifically than “women, lower-income, and the youngest and oldest age segments” - who are the people who aren’t using Wikipedia as much, or not at all, and why not? What can we do to get them to use us, or use us more?
Stage 1: Validate and refine list of Wikipedia use cases
Approach
- Quicksurvey of ~200 in-the-moment readers
- Invite them to click out to an external survey ( hosted on Qualtrics)
- A small number of open-ended questions.
Stage 2: Large-scale QuickSurvey of on-Wiki readers
Approach
- QuickSurvey of several thousand in-the-moment readers
- 3-5 questions
- Use case(s)
- Demographic information
- Opt-in to Stage 3 (in-depth survey)
Stage 3: In-depth survey of readers
Approach
- A longer survey focusing on areas to include:
- Motivations/goals/tasks that take you to WP
- Session behavior questions
- Motivations/goals/tasks that take you to AI tools and associated other knowledge platforms, e.g.:
- video platforms like TikTok and YouTube,
- social media platforms like Facebook, X, and Instagram,
- messaging platforms like WhatsApp,
- educational platforms like Coursera,
- search engines like Google, and
- community forums like Reddit, Quora, and X,
- Frequencies of visiting Wikipedia vs other platforms
- For [these use cases], what platform(s) do you go to?
- Trust measures for both WP and AI tools/other knowledge platforms
- Hypothetical scenarios: where would you go if this happened to you?
- Pain points associated with individual use cases
Stage 4: Diary study and in-depth interviews
Target population
- External panel participants with a range of Wikipedia reading frequencies and Wikipedia-engagement levels
- Design Research database participants who indicate that they are engaged power readers
Approach
- A daily survey (distribution mechanism TBD)
- Subset of participants invited to post-completion interviews
- Specific focus on:
- The Wikipedia reading experience
- What is a good reading experience?
- How personalized, lengthy, or source-based do participants want their sources of information to be?
- How much of their information is transmitted via text?
- Participants’ informational/learning landscapes
Timeline
editStage 1 is currently underway, and we hope to have all stages completed by the end of Q2 (December, 2025).