More Answers For Less Money Can Leave the Platform Poorer

Based on the research ofYi Gao, Amit Mehra, and Dengpan Liu, "More Can Be Less: The Economics of Answer Viewing on Paid Q&A Platforms," Information Systems Research, 2026

Paid Q&A platforms built their business on a simple exchange: a user pays, an expert answers, the platform keeps a commission. Over the past several years, a number of these platforms have layered on a second product, often called answer viewing, that lets a different user pay a smaller fee to read an answer that was already written for someone else. It looks like nearly free incremental revenue. The question already exists, the answer already exists, and the platform just found a new way to sell something it already owns. A new paper in *Information Systems Research* by Yi Gao, Amit Mehra, and Dengpan Liu takes that intuition seriously enough to model it formally, and the model says the intuition is wrong more often than platform operators might expect.

The Feature That Looks Like Free Money

Gao, Mehra, and Liu build a game-theoretic model of a paid Q&A platform with three parties: the platform, the answerers who set consulting fees for personalized responses, and the users who decide how to get their questions answered. Before the answer-viewing feature exists, a user with a question has one real option: pay an answerer's consulting fee for a personalized response. After the feature launches, that same user has a second option, paying a smaller fee to see whether someone already asked something close enough to be useful.

It is worth being precise about what kind of paper this is. This is not a field experiment or an analysis of platform transaction logs. It is a theoretical model: the authors set up the incentives of platform, answerers, and users as a formal game and solve for how each party would rationally respond to the introduction of answer viewing. That matters for how much weight to put on the findings. A model like this tells you the logical structure of the problem, which effects push in which direction and under what conditions, rather than a measured effect size from a specific platform's data. The value of the paper is that it isolates a mechanism that is easy to miss when you're staring at a single quarter's viewing-feature revenue line and calling it a win.

Why Cheap Access Cannibalizes the Expensive Kind

The mechanism the paper isolates is cannibalization, and it runs through substitution rather than expansion. The optimistic story for answer viewing is that it opens the platform to a new segment of price-sensitive users who would never have paid for a personalized answer, so the feature is additive. The paper's model shows that story is only half true. Some of the users who buy access to a viewed answer are indeed new demand the platform wouldn't have captured otherwise. But others are users who would have paid the full consulting fee for a personalized answer, and now choose the cheaper substitute instead. Every dollar captured from that second group is a dollar the platform used to get at a higher price, now captured at a lower one, net of whatever cut goes to the answerer whose original response is being resold.

Whether the platform ends up ahead depends on the balance between those two groups, and the model shows there is no guarantee it tips the right way. The platform can end up worse off from adding the feature, precisely because the users most likely to actually use a cheap viewing option are disproportionately the same users who were already profitable customers of the personalized-answer market. A feature aimed at the users a platform doesn't have ends up mostly serving the users it already had, at a lower price.

A feature built to expand the market can end up mostly reselling the customers a platform already had, at a lower price than they were already paying.

The Revenue-Share Trap

The paper's sharpest finding is not the cannibalization risk itself, which is a familiar concern in platform pricing, but what happens to answerer behavior once the platform decides how to split viewing revenue with the experts whose answers are being resold. The intuitive assumption is that giving answerers a larger cut of viewing revenue should make them happy and leave their consulting fees for personalized answers unaffected, since it is simply extra income layered on top of the existing business.

The model finds the opposite can occur. When the platform shares a large enough portion of viewing revenue with answerers, answerers respond by raising the consulting fee they charge for personalized answers. The logic is that the answerer is now earning meaningfully from the passive resale of past answers, which changes the value of continuing to sell new personalized answers relative to that passive income; the answerer's calculation for what a personalized answer is worth shifts, and the new equilibrium consulting fee is higher, not the same. That has a direct knock-on effect for users: the group still paying for personalized answers, often the users with the least generic or least previously-answered questions, ends up paying more for the exact same service they could get before the feature existed. The paper's authors describe this as a case where users do not always benefit even when the platform's overall numbers look fine, because the users who remain in the personalized-answer market are the ones absorbing the fee increase.

What This Looks Like Outside the Model

The paper doesn't name specific platforms, and no company's own data is used to test the model, but the pricing structures it examines already exist in the wild, and looking at them shows why the tradeoff is a live design question rather than a hypothetical.

JustAnswer, a paid Q&A service founded by Andy Kurtzig in 2003 that connects users with roughly 12,000 verified experts across categories like medicine, law, and home repair, already runs a version of the tiered structure the paper models: a user can pay a per-question fee ranging from about $5 to $90, or join as a member and pay a flat monthly fee for unlimited questions instead. On the answerer side, JustAnswer already varies its revenue share: a new expert starts at 20 percent of what the user pays and can work up to 50 percent as ratings accumulate. That isn't the same mechanism the paper models, JustAnswer's share scales with reputation rather than a resold-answer viewing product, but it shows paid Q&A platforms are already comfortable treating the expert's cut as a behavior-shaping lever, exactly the lever the model says has to be calibrated carefully.

Quora runs a structurally similar experiment on the creator-payment side. In an official 2020 announcement from CEO Adam D'Angelo, Quora introduced Space subscriptions, letting any creator set a subscription price and keep a 95 percent share, with Quora taking a 5 percent platform fee, alongside a separate Quora+ bundle that distributes subscriber payments to creators in proportion to how much of their content a subscriber consumes. Neither product is literally a fee to view one prior answer, but both are instances of a platform choosing how much of a new monetization layer to pass through to the people who created the underlying content, the same design choice the paper shows can move consulting-fee behavior.

The starkest cautionary tale, even though it involves a different kind of cheap substitute, comes from Chegg. Chegg built a subscription business substantially around 24/7 access to a network it has described as more than 150,000 subject-matter experts answering students' questions. In its first-quarter 2023 earnings release, Chegg reported total net revenue of $187.6 million, down 7 percent year over year, and 5.1 million subscription subscribers, down 5 percent year over year, with CEO Dan Rosensweig explicitly framing the company's strategy around responding to "artificial intelligence technology" in that release. Chegg's problem wasn't an internal viewing tier cannibalizing its own personalized-answer product; it was an external, far cheaper substitute, a free chatbot, pulling users away in a pattern that echoes what the model predicts for an internal cheap alternative. The lesson generalizes: a business built on personalized paid answers is exposed the moment a cheaper way to get an answer exists, whether the platform builds that option itself or someone else builds it for them.

The Rule Worth Taking From a Model, Not Just Data

The honest caveat is that this paper is theory, not a natural experiment inside a real platform's transaction logs. What it offers instead is a clean account of why a monetization feature that looks purely additive can behave differently once users substitute and answerers reprice in response. For any operator eyeing a similar feature, the implication is that the viewing price and the revenue share to the original answerer aren't independent dials to set once and forget. They interact with each other and with the price of the core product in ways that can turn a feature meant to expand the market into one that quietly shrinks it.