Many companies today uses Google Suite for their Email and everyone has gotten used to the Gmail interface. One advantage of Gmail is, safe and easy programmatic access to the emails via IMAP.
The problem I am trying to solve here is you might not know when a person would respond back when you send an email. You will compose an email, hit send and wait forever. I know that Emails are Asynchronous, but it would be a nice to have an idea of when one can expect a reply.
How it will look like
The product is a chrome extension which will be invisible 99% of the time, unless you are composing an email. When you are composing an email or replying to an existing email, it will see who the recepients are and will show a single line notification saying when you can expect a reply.
A simple sentence like “You can expect a reply from John Doe in about 2 hours (based on past history)” would be enough. If there are multiple recipients, it will show a table with a row for each user.
How it works behind the scenes
This is a service that you will sign in to and allow programmatic access to all your emails. For the privacy conscious, this service doesn’t have to read through your emails. It just needs the sender, recipients and the date and time of the emails.
Using all this data, it needs to build a predictive model for each user which gives the time when one can expect a response. The different factors to take into consideration are the sender, other recipients in to/cc, date and time of the email, day of the week, etc.
If the user doesn’t care a program reading through the contents of the email, even the subjects and body might be used to predict better. Maybe some people would respond to specific subjected emails and would ignore other general marketing emails.
As more and more people use the service, you can start building a nice prediction graph of how the user responds to other user’s emails. This would work well in a Google Suite Gmail app where everyone in a company are signed up to this service.
The extension will also suggest the right time to send an email to a list of recipients so that it is viewed and responded immediately. It can also handle auto sending at the optimum time to make it all automatic.
It would be very useful for sales people and marketing teams to send the most effective emails at the right time. I think project managers will also find it useful as they would like to know when to expect a reply from their direct reports.
Coming up with the right pricing plan is important. Since this tool would only work if lot of people are using the service, it is important to have a free plan, but with limited feature set. Maybe you can only get approximate time of reponse with hour granularity.
The paid plans can have even more finer granularity (at minute level) and also utilises the insights it learns from others. By paying for the service the users can also make use of the automated sending at the best time feature.
This is just a simple idea I had noted down in my Evernote a few months back. I haven’t thought of how to implement the machine learning pieces or how to scale the prediction models at millions of users scale. If someone is interested in building this, do leave a comment below and I would like to help flesh out the details.
Also published on Medium.