Launch HN: UPchieve (YC W21 Nonprofit) – Live tutoring for low-income students
The idea for UPchieve was based on my own personal experience as a low-income student. I was raised by a single mom who was an immigrant to the US and often couldn’t help me with schoolwork and applying to college. So I went to a community college, and from there it took me 6 years to transfer to and graduate from a 4-year university.
College is not the only way for a low-income student to achieve upward mobility, but research shows it’s one of the most reliable. That was the story for me, too. Transferring to Penn helped me get a job on the trading floor at J.P. Morgan, where I immediately made 3x as much as my mom straight out of college.
I soon began looking for ways to give back and help other students like me get to college. Volunteering as a math tutor seemed like a great choice because I had tutored my way through HS and college and knew a lot about it, but I couldn’t find any existing volunteer opportunities that fit with my work schedule.
That’s when I first came up with the idea for UPchieve: an open source platform that could meet the needs of both younger-student-me (who needed help with schoolwork and college apps, often late at night in my home) and older-volunteer-me (who needed a flexible and convenient way to give back). Students can use UPchieve 24/7 to request and get paired with a live tutor in 5 mins, and volunteers select their availability and then wait for a text notifying them that a student needs their help. You can see a demo of both the student and volunteer side of the platform here: https://www.youtube.com/watch?v=SufRUje0XiM
Over the next two years, I realized that many other people cared about what I was doing and wanted to volunteer to help build the platform and tutor students. I also learned more about the state of education in the US and discovered that the gap in college completion rates between low- and high-income students hadn’t changed at all in 10+ years! This led me to grow more and more passionate about both the need for and potential of a platform like UPchieve. In 2018, I finally quit my job at J.P. Morgan to work on UPchieve full-time, with the new goal of scaling free tutoring to all low-income HS students across the US.
UPchieve is completely free for eligible students, and it only costs us $10 to give one student an entire year of unlimited academic support. For reference: online tutoring companies charge students $35 per hour on average. Other existing nonprofit tutoring models are prohibitively expensive and difficult to scale, due to factors like 1) cost of paid tutors, 2) in-person delivery, and 3) long-term commitments required of students and tutors.
We’re also working to become 100% self-sustaining (i.e., not dependent on donations). Companies like Verizon and Goldman Sachs pay us to provide virtual volunteer opportunities to their employees. Many big companies have budgets set aside specifically for volunteer opportunities because studies have found that employees who volunteer are more likely to stay at the company.
So far, we’ve helped more than 7,000 students and matched more than 30K on-demand requests for tutoring. My favorite part of my job has been reading the hundreds of comments we get from students each month, telling us how we’re helping them kick ass in their classes and that they can’t imagine life without us. It’s literally the best feeling ever. You can see some of my favorite student feedback here: https://docs.google.com&#...
66 comments
[ 2.4 ms ] story [ 1153 ms ] threadSupply is definitely going to be a challenge for us in the long term. We think we’ll need over a million volunteers someday to support all 8 million low-income HS students in the US. However, at least during COVID-19, we found it really easy to recruit volunteers because everyone was looking for virtual opportunities. At one point we amassed a wait list of over 10K people who wanted to volunteer with us.
I think one thing we have going in our favor in the long run is that we designed the volunteer opportunity to be convenient enough that even busy professionals should be able to take part. Volunteers have complete control over their own schedule and how much time they want to work with students in any given week/month/year.
One study found that 50% of Americans say that they want to give back but can’t because of time and schedule constraints. About a third of Americans are college-educated and likely able to tutor at least one subject, so that puts us at about 50M potential volunteers just in the US who might be both qualified and interested in our volunteer opportunity.
Through corporate partnerships with big companies and potentially even clubs on college campuses, we’re hopeful we can recruit enough volunteers to meet student demand in perpetuity. And to address your point about recruiting for harder subjects… another thing we have in our favor is that the distribution of individuals qualified to help with different subjects matches up pretty well with the distribution of students needing help in different subjects. Our most popular subjects on the platform are actually lower-level math subjects like Algebra 1 and Geometry, and those are also the subjects volunteers are most likely to get certified in.
Recruiting retired teachers could be a good option for us too since they have the expertise of a highly trained teacher without the cost. At a minimum, I agree we need to invest even more into the training of the volunteers!
How do you stop predatory paedophiles from abusing young people using your service?
> Our coaches will never make you feel bad to ask questions—even the ones you’re too embarrassed to ask in class.
How do you ensure this?
It does look great though.
We try to prevent abuse through a few measures:
1) The platform is only available to students who are 13+ , and we encourage both students and volunteers to keep their conversations focused on academics. When in a session, students and volunteers only see each other’s first names (which is not a lot to go off of, considering we have users across the country and even some international volunteers).
2) Before volunteers can begin working with students, they have to go through our screening process. Volunteers provide a picture of their photo ID and two references who we contact and ask about their suitability to work with children. A human being reviews all reference forms and runs the name from their ID through the National Sex Offenders Registry.
3) Tutoring sessions on our platform are completed using a whiteboard and text-based chat (no audio and video). We have a chat filter that prevents inadvertent attempts to exchange personal information like phone numbers and emails, and we also block links out to common third-party video conferencing sites like Zoom.
4) Students and volunteers can rate and report each other, and we read through the chat logs of most of the sessions on our platform. Sessions can get flagged for review for tons of different reasons. This is the piece of our process that is least scalable, and we’re still looking for ways to balance quality assurance and safety with scalability.
Getting back to the safety aspect - people are less likely to test the boundaries of your safety rules if they know these (anonymized) sessions will be seen by many other people, and this would massively increase the number of eyeballs for each session further increasing the detection rate of any remaining bad behavior
[Edit - you could also use these recorded sessions to help volunteers get better at teaching, you could get an expert to put together small videos on improving teaching technique with positive and negative anonymized examples from the sessions, then make them available to the volunteer tutors. That could be another low cost and scalable way to reward the tutors - most people who teach would like to get better at it!]
One thing we probably won't do any time soon is make a library of past sessions available publicly to students. We want to stay firmly in the category of live, personalized support so that we don't overlap with sites like Khan Academy (videos) or Quora/Stack Overflow (asynchronous support).
We do really want to expand our volunteer training and get volunteers to provide feedback to their peers on their sessions though, and I agree the recorded sessions would help a lot with that!
Have you found that to be sufficient? I recently tried doing some online tutoring, and I found that even with an audio connection, I had a lot of trouble telling if anything I said was landing at all. I came away from the experience rather disillusioned with the idea of trying to do this online at all.
On our site, students seem to really like the text-based communication because it creates what we call a "judgment-free zone". They have time to think through what they want to say and can end up less embarrassed to admit when they don't understand something.
On the tutor side... yes, it does require the tutors to be particularly engaging and patient. :) Our best tutors try to gauge students' understanding often and ask lots of questions to elicit active participation. Like anything, it just gets easier with practice!
What's wrong with just enabling audio?
In the future, we actually do plan to add audio to the platform as an optional feature that students can turn on. Most students prefer the text-based communication of our current platform, but there are definitely some who would use audio if given the choice. We still need to think through how to get mutual buy-in from both the student and volunteer in that session as well as whether we want to do additional background screening for volunteers who want to use that feature.
Do you see yourselves taking a hybrid approach to this to optimize on that spectrum? Can someone get involved if their schedule is too unpredictable to do anything but the former? As a startup cofounder who would love to teach, I’m very much asking for myself :)
To your excellent point about continuity, relationships are definitely a component of effective tutoring, but in our user research and conversations with students, students themselves told us it was more important to be able to get help when they need it than to be able to get help from the same person every time, and that's the need we've decided to optimize for.
Is there some mechanism by which volunteers can find an appreciative student and re-match with them exclusively?
It's really common for students to thank their tutors at the end of the session and leave comments in the post-session feedback form saying how awesome the tutor was. Right now we're just sharing that feedback with the volunteers ad hoc, but we'd like to find a better system to share it regularly.
We're also planning to launch a volunteer favoriting feature for students later this year, which will increase the likelihood of students and volunteers who get along well having future sessions with each other. It's actually a commonly requested feature from both students and volunteers!
I wonder if a 'karma' system is possible by which students and tutors could give positive feedback to one another and students could 'unlock' highly rated tutors and visa versa. As someone who has spent a fair amount of time tutoring students (when I was a student myself, admittedly) I can certainly say that tutoring an unappreciative student will quickly sour the experience for a lot of folks, even if the majority are appreciative.
And hopefully we volunteers as a whole can make the platform feel like a partner that students can continuously rely on!
It wouldn't have to be just for tutoring, but it can help in terms of mentorship and just learning from older people with more real life experience and career.
Other thoughts: - I thought the signup buttons/some action buttons could use a color contrast - some of them were hard to see visually. - Some of the visual indicators when the mouse hovers over a button/action would be nice, just so the user knows which items are more clickable - more affordances can be nice. - I put in my preferred name to be called during sign up.IF you are going to do verification, maybe you can make that clear. Or have the option to keep a nickname. - The profile page needs more information for editing and display. But I think it is probably on your backlog.
The problem of how to get revenue, not just donations for something like this is an interesting one. One idea it might be worth exploring is charging colleges in order for them to be able to contact students through the site as part of their efforts to increase diversity (It would need to be opt-in for the students, to maintain their privacy). If you can keep students on the platform through college, you could do the same for employers looking to increase diversity in hiring - maybe the way to do that is encourage students to “graduate” to being volunteers on the platform when they finish high school? Then the (college) students would get the benefit of support and outreach from employers for internships during college, as well as the psychological and learning benefits of tutoring others who are a few years behind them.
I think it would also be interesting to consider making some kind of private social network for your student “alumni” - having a peer group can be super helpful for low-income students in college, so this would hopefully keep them involved with upchieve, make them more likely to succeed in college, and hopefully let you charge companies looking to recruit them.
Happy to chat about this or any other crazy ideas, contact details in my profile
Also, you mentioned it’s open source, link to the source?
Student side is currently only open to low-income high school students in the US who are 13+. Eligibility is based on a combination of the income demographics of their high school and zip code, but even students who don't pass the "quick check" can apply for access if they feel they are eligible.
https://gitlab.com/upchieve
1) Expand to offer more subjects. Students ask us all the time for tutoring in English and history. Someday we hope to offer tutoring in every common high school course, including foreign languages and computer science too!
2) Improve the reliability of the site. We have something like 100x as many tutoring sessions happening every day as we did last year, and we need to make substantial changes to the platform to keep up with the usage.
3) Upskill our volunteers to improve the quality of tutoring offered on the site. We want to offer additional training courses on tutoring skills and implement a peer feedback system so volunteers can help each other improve.
4) Reach more students! We want to get the word out to as many students as possible about our free service. If anyone reading can make an intro to a high school or nonprofit who needs us, we'd greatly appreciate your help.
Our long-term goal is to have engaging, fun feature work for community contributors to work on, and we do have those come up occasionally now, but we can use a lot of help just finishing our conversion of the backend to TypeScript, updating dependencies, small bug fixes and similar work, that we have issues ready for.
We also have a weekly tech community Zoom meeting that I'd be happy to send an invite for to anyone who emails me with interest. dave.sudia at upchieve dot org
Best of luck scaling to support 100x users! In case you are not already using it, New Relic's new program for nonprofits provides their entire platform for free, and it can be really helpful for scaling by seeing how the app is performing. In case you are not already using it, here's the link: https://newrelic.com/blog/nerd-life/introducing-observabilit...
I tutored math for extra cash in grad school, and I was almost exclusively hired by well-off local families so that their their high schoolers could improve from an A-minus to an A, ace the AP Test, test into Calculus 2 at UMich, and have a leg up getting into the business school. I loved talking about calculus every day, but it felt like I was participating in a weird arms race between these parents, and upholding a status quo that I really don't agree with.
Tutoring other students in undergrad was much more fun and rewarding -- even though the pay was 1/6 the rate -- because it was paid for by the university and anyone could walk in and get help.
I hope this turns out to be sustainable, as I'd love to give some of my time to it.
I also did a lot of paid tutoring for more privileged kids in HS and college, so I can relate. Fortunately that experience ended up being very useful in starting UPchieve!
Based on your past experiences, it sounds like you'd be an awesome volunteer! If you want to just learn more about the volunteer experience you can check out this page on our site: https://upchieve.org/volunteer
Since we're open source, I think it'd actually be awesome if someone in another country decided to start their own version of UPchieve for that country!
Can I ask about safe-guarding? It's not the first thing you think about but if you cast your mind back to Airbnb or Uber, their biggest issues were safety.
I would imagine background checks would be insufficient - do you plan to monitor / record sessions? What training is in place for the 30k volunteers? Will there be mentors / chaperones who can drop in on calls and check?
Even with the best of intentions a middle aged guy trying to teach statistics for two hours a week to a 16 year old girl is going to be awkward. Without the best of intentions that's a minefield.
That said - hope you blow the doors off!
The scalability our volunteer training effort is definitely something we're actively working on figuring out. We're trying out peer mentoring between volunteers and that has gone really well.
It's also a major reason we are chat/whiteboard only.
If a student does not automatically qualify, we offer them basically an appeal form where they can say how much their family makes among other things. And we do qualify a number of students that way.
Partnering with other orgs like BGCA is an awesome idea that I will pass to our outreach coordinator. Thank you!
I grew up on the bottom rung of our local socioeconomic ladder, and self-selected out of a lot of potential opportunities because I knew I would be excluded.
We try to avoid using the word "low-income" in all of our student-facing communications for exactly this reason. For example, this is how we talk about eligibility on the students' page of our website (https://upchieve.org/students):
"Is This Really Free? YES! UPchieve is a nonprofit organization that exists to help students achieve their academic goals. Any student whose family can’t afford a private tutor or admissions consultant is eligible for free tutoring and college counseling through UPchieve. Instead of charging you, we rely on volunteers and donations to provide you with free academic support."
One of my long-term goals with UPchieve is actually to destigmatize being a low-income student. It's obviously not the student's fault if their family can't afford something. (In fact, it's probably not their parents' fault either.) Not sure exactly how we'll do this yet, but I hope that we can :)
Also, are there any plans to offer non-English services? A lot of recent immigrant families would benefit from tutors who are bilingual in their language.
Going from a primarily low-income school to a good college showed me how much of the achievement gap in college between poor and wealthy students simply comes down to tutelage. This kind of work is really important - kudos to your team.
Re: tutoring in other languages, I completely agree it's needed. We hope we can add tutoring in at least Spanish in the next two years. We've already started asking volunteers during onboarding what other languages they speak, but as of now we don't have enough volunteers fluent in other languages to launch tutoring in those languages.
For the upchieve101 quiz - some of the questions weren't covered in the training (or I missed it e.g. what subjects do students want most help with?).
The implicit bias stuff was annoying - I agree with the idea, but there is little evidence it makes any difference to behavior: https://www.scientificamerican.com/article/the-problem-with-...
Having to implicitly agree with ideas like "being colorblind is racism" (paraphrasing) was almost enough to put me off continuing - and being told that it was a sign of white racism when I'm a brown immigrant was a little weird and suggested some unhelpful assumptions on the part of the material's author(s)
It was nice to help a student, and I think they learned a lot during the session, but I was surprised how weird it felt not having any audio interaction. Everything took much longer than it would have otherwise.
I understand the need for student safety, but not knowing anything about a student, and knowing I would never know how this student would do in future was also demotivating, it felt like taking part in "mechanical turk for tutoring". I'll try a few more times, maybe I'll get used to it. But rich kids somehow manage to stay safe with online tutors with audio and video, seems a shame to give the poor kids a crappy version of the same thing.