Our workflow is evolving fast. We started with a small team trying to cope with a great demand; then we adapted to be able to grow the team and slice the work in small pieces that can be handled in a distributed way; now we are pushing to grow even more and improve the quality and speed of our work. What follows may therefore be short-lived, but this is how we are working right now.
We are currently receiving 5-10 requests per day, but only half of these will pass our validation process. The other half are requests that seem unfeasible for us, because they are too broad or out of our scope. Once a request passes validation, we write it up for the task distribution form. This writing usually starts by copy-pasting the original request, but the whole team discusses about each request, trying to figure out whether it’s clear enough or some key information is missing. While this is hard for very technical tasks, now we have a big enough team so someone is familiar with almost every topic. Once we are happy with the request description, we mark it as “Ready for task distribution e-mail“.
Roughly every two days we e-mail all our volunteers a form with all our new requests, so that they self-assess their ability to help with each request. The e-mail is now reaching about 35 000 volunteers (in total >45 000 volunteers subscribed to the platform, but only 35 000 chose to receive our e-mails). We estimate that about half of these 35 000 volunteers open the e-mail, and about a quarter of them click the link to the form. This is more than 8000 people looking at the requests, but only half of them actually click the “submit” button at the end. All in all, 24 hours after sending the e-mail we collect between 4 000 and 5 000 responses.
Given that we send task distribution e-mails every two days, up to 48 hours may pass between the moment in which a COVID-19 researcher makes the request and the moment we send it to the volunteers. Then we need to wait another 24 hours for enough volunteers to respond to our form. This is the slowest step of our workflow, and we’re working to improve it. One obvious way would be to send the e-mail every day, but we’re worried to overload our volunteers. Their attention is our most valuable asset.
For each request, one of our advisors checks all the feedback from the volunteers in order to find the best match. For generic tasks we may have hundreds of volunteers with great skills. For very difficult and specialized tasks we may have none, but often some volunteer knows someone who can help (or knows someone who knows someone). In most cases, even for highly specialized tasks we still have several amazingly skilled volunteers. Out of the 8 000 people who read our task distribution form, sometimes only one is the perfect match.
A coordinator takes the recommendations from the advisor and sends a proposal to the COVID-19 researcher. At this time we put the last green square in our progress table, but the work is not done. We still follow up and do all we can to reach actual success for the task.
Success is a heavy word for us. We want to make sure that we have a true impact, so trying is not enough. If a researcher needs a reagent, we are not satisfied sending a link to a provider or the e-mail of a postdoc who has it. Our work only finishes when the reagent arrives. This is a hard criterion and often an unfair one. Many hours of hard work go to the “failure” column for requests that were in fact impossible. But we are happy with this criterion. Each success is the distilled product of a long line of hard work and good will, and each success is worth it.