"They were not able to meet the promises of the service." That was the verdict of a New Mexico superintendent on the tutoring company his district tried for one year and then dropped. The company later closed the school service altogether.
It is a story about school contracts, but it is also a story for any parent choosing a tutor, whether human or AI, because the reasons it went wrong are the same questions worth asking before you sign up for anything.
What happened?
Varsity Tutors for Schools, one of the largest virtual tutoring providers in the country, told its school customers at the end of July 2026 that it was discontinuing the service, effective immediately, according to reporting from EdWeek Market Brief (EdWeek Market Brief) and The 74 (The 74). In a shareholder letter, the parent company, Nerdy, said it would wind down the online school service it had launched during the pandemic and focus on its original business, selling directly to parents and college students, where "the opportunity is significantly larger."
It was not the first. Another large provider, FEV Tutor, had terminated abruptly in early 2025 as federal relief money ran out.
Was it just about money?
Partly. Matthew Kraft, a Brown University professor who has studied tutoring, told EdWeek Market Brief that federal recovery funding gave districts both the resources and the urgency to expand tutoring after the pandemic, and that a whole marketplace of private providers grew to meet the demand. Then the temporary money ended. But The 74 reports that several states, including Louisiana and Massachusetts, now provide ongoing funding for high-impact tutoring, and that districts have become more demanding about results.
Tutoring experts quoted by The 74 said Varsity Tutors for Schools was not able to deliver proof that its services raised student performance, though the company's own explanation was a refocus on its consumer business. Kevin Huffman, CEO of Accelerate, a nonprofit that funded Varsity last year as part of a research study, wrote that some tutoring companies "pushed (or were pushed by funders) into rapid growth ahead of evidence-building" and were "now paying the piper." Other programs, he noted, had pushed for strong external evaluations, including randomised controlled trials.
What did a school district see?
Matt Montaño, superintendent of the Bernalillo Public Schools in New Mexico, chose Varsity in 2024 when the state education department struggled to roll out a statewide tutoring program. The company had promised to "immediately" connect students with a tutor, day or night, whenever they needed help. He told The 74 that Varsity was unable to provide support "in a timely fashion," and that his district stopped working with the company after one year.
Notice the gap between a promise and the daily experience. A student who needs help at 8 p.m. on a Tuesday does not care about the pitch. They care whether a tutor appears.
What does good tutoring look like, according to the research?
The article also describes what the evidence does support. Research favours models in which students meet at least three times a week, one-on-one or in small groups, with the same tutor over several weeks. Kathy Bendheim of the National Student Support Accelerator, a tutoring research center at Stanford, said the providers most likely to have stronger outcomes are those that "adhere closely to the research." That is a useful test for any service you consider: does it follow the pattern that has been shown to work, or does it simply offer availability?
How should you read a company's own results?
The 74 notes that Nerdy's website includes case studies. One from 2022 describes a small study in California's Garden Grove district, where fourth graders at or below grade level in maths were randomly chosen for after-school sessions with a Varsity tutor. According to the summary, their i-Ready scores rose by 9.82 points, against an average of 4.52 for students who did not get the tutoring. That looks like a positive result, and it may be. But it is a summary on a vendor's own site, from a small study several years old, which is the kind of evidence to treat as a starting point for questions, not as proof.
What can parents ask before choosing any tutor?
Whether the tutor is a person, a program or a mix, five questions come straight out of this story. Ask what evidence exists that the service improves results, and who gathered it. Ask how often your child will meet with the same tutor, since the research points to regular sessions with the same person. Ask what happens if help is not available when your child needs it, and how the service handles a session that goes badly. Ask what data it keeps about your child and what happens to it if the company closes. And ask how you will know within a month whether it is working.
Our guide to choosing between an AI tutor, a human tutor and Khan Academy walks through those choices in more detail, and our eight questions to ask before choosing an AI tutor is a checklist you can take to a demo.
Frequently asked questions
Does this mean online tutoring does not work? No. The research the article cites supports tutoring that is frequent, small-group or one-on-one, and with the same tutor. Experts said this provider could not show evidence of results, and one district reported that it did not meet its promises.
Is my child affected if this company was their tutor? Varsity Tutors' consumer business for parents and college students continues, according to the shareholder letter. The school service was the part that closed.
Should I only choose tutors with published studies? Not necessarily, but ask what evidence they have, and treat vendor summaries as a starting point. Independent evaluations are stronger evidence.
What about free options? Free resources are a sensible complement. A well-used free platform can beat an expensive service your child rarely uses.