automating drop-in tutoring

Sep 6 2024

Automating Drop-in Tutoring

Context

I’ve been a Writing Fellow for 3 years. We are undergraduate peer tutors who work with students to strengthen their writing abilities. Fellows help students with

all types of writing projects (lab reports → graduate school essays)
at any stage of the writing process (brainstorming → final polishing)
with any writing concerns (grammar → organization of ideas)

Side note: I love my job. I love seeing writers becoming more confident as their writing takes shape. I can talk about it all day :-) As Writing Fellows President, I get to help other tutors do what they love better.

In addition to scheduled tutoring (by appointment), the Writing Fellows provide drop-in tutoring. Only one drop-in tutor is available every hour, so when multiple students want to drop-in, we implement a waitlist:

This waitlist is first-come-first-served.
This waitlist does not carry over to the next day.
Once a drop-in tutor contacts a student, an appointment will be reserved for a duration of 10 minutes.

What's Wrong?

Previously, drop-in fellows are encouraged to check the drop-in waitlist periodically throughout their shift, which means repeating (part of/the entirety of) the following workflow multiple times every hour:

Previous Drop-in Workflow

Sounds simple enough? Nope! This is actually very time-consuming and error-prone for the drop-in tutor. Specifically,

Manually checking for new responses leads to errors. Reading 1 row of a Google Sheet is easy. Reading 40 rows, all populated with information, is not. Having to skim the entire populated sheet for new responses, then read the contents of each row to identify appropriate students means a lot of room for error. Tutors can misread. If this happens, they could fail to contact students who need help.
Interacting with multiple platforms slows us down. Try switching between viewing the sheet on your laptop and texting on your phone! Or worse yet, switching between everything on your phone. I get chills just thinking about it.
Editing the form-populated response sheet creates data security concerns. Tutors may accidentally delete form responses, incorrectly annotate rows, etc.

Previous Waitlist sheet is populated by Google Forms, but is also edited by tutors

To the Rescue...

I optimized the drop-in waitlist handling process to address all of these concerns.

Previous Drop-in Workflow

1: Automate!

Eliminate user error by automating checking for responses. I wrote a Google Apps Script so that tutors get sent an email whenever there’s a new response. Instead of parsing through the populated Response sheet, tutors only need to read the contents of the email – equivalent to reading only 1 row of the sheet!

What’s more, only the right tutors get contacted whenever there’s a drop-in request.

Emailing all 40 Writing Fellows every single time a student signs up for drop-in tutoring is not ideal. After all, the drop-in request is only applicable to the day it is submitted, and only drop-in tutors are concerned. This switch statement makes it so that only the drop-in tutors for the specific day are contacted.

2: Speed up!

Speed up tutors' communication by consolidating interactions. I configured the email that’s sent to drop-in tutors so that they can directly copy & paste information on their phone. This eliminates the need to search for a student's phone number or for the text template.

3: Secure Data!

Ensures data security by modularizing information sources. The previous sheet-editing procedure muddled two actions: gathering Google Form responses and record-keeping for tutors. We need to separate these concerns. The sheet with responses populated by Google Form should be view-only.

I created a separate editable sheet, synced with the form responses, and instructed tutors to edit the sheet for record-keeping only.

I restructured the editing permissions of the Form-populated responses sheet accordingly.

Impact

For this system, I have to change 1 line of code whenever tutors cover each other’s shift(s) (~30 seconds). This is negligible considering the costs saved for the Writing Center. With this, I’m saving:

5 minutes (of manual checking, etc.) / drop-in shift

* 26 drop-in shifts / week

* 12 weeks / semester

= 26 hours every semester

That’s $400 worth of redundant work every semester...

without including the time/$ costs of errors – contacting the wrong students, recovering accidentally deleted data, and most egregious of all, missed tutees. The impact of delivering the right help at the right time can’t be quantified in numbers - it's invaluable. This process automation helps deliver that unquantifiable value :-)