How would you improve the Drafts feature of your mail client?

Candidate: I assume that the goal is to make it easier for the customers to getting things done and increase their productivity through enhancing this draft feature.

Interviewer: Sounds good

Candidate: Do we want to improve the back-end experience such as latency and performance, or the front-end experience

Interviewer: Front-end experience

Candidate: Do we plan to roll out the improvements to app and desktop or both

Interviewer: Both

User segments

A) Personal users

B) Business users

To maximize the reach and impact, I would focus on the personal users that are larger in size

Mapout the user journey

A) Start to write a mail

B) Need to take a break, save the content as draft

C) Retrieve the mail from draft and start from where it was left

D) Complete the mail and hit send

Identify the pain points

A) When the battery dies out, auto-saved drafts does not capture the entire content up until that point

B) Users tend to leave the drafts incomplete forever

C) Users forget about drafts, start to create a mail content from the stratch again

Brainstorm the solutions

A) Intelligent auto-save: When the battery is running low, start to do a auto-save every second

B) i) Snooze: Create tools that users can use to set a reminder when they want to work on the draft again.

ii) Auto-remainder: Based on the content and key words and receipients on the draft saved, create a auto-nudge for the users to complete

C) Content matching: When users start to a write a mail, check if the mail content matches with the draft content, in which case, show the pop-up panel asking "reuse the draft content". User can re-use the draft with one-click approval.

Prioritization

Solution	Effort	Impact
Auto-save	Moderate, create a module that takes a auto-backup	High, uses do not lose there work
Snooze	Low, simple remainder feature	Moderate, not all users would adopt the feature and not all uses act when snooze goes off
Auto-remainder	High, need to built advance personalization ML models to identify the ones that are important	High, users do not miss out on completing the important mails
Content-matching	Moderate, need to build matching models	High, saves time for users

If I were to prioritize one feature, I would go with content-matching, because users save time and no other competitors currently offer this feature for the users.

Code Uncode · Answer 2 · 2021-12-05T05:30:04+0000

Clarifying questions:

What is the driver behind this need to improve the draft feature?
Is it driven by member feedback, client request, market research or is it a response to a feature a competitor launched?
Is this a feature that we want to deploy on the web app, the phone app, or on desktop clients?
For most people, drafts are simply an email that is not sent yet, however, some users use them to keep to-do lists, self-notes, templates, or emails that they are not sure they want to send immediately. The most appealing aspect of a draft is its simplicity; however, we know that very often features are stealthily launched and people do not even know they exist. Do we have to build a new feature, or simply design a way to bubble up an already existent feature, or plugin something a feature from a different solution?
When I think of email, I broadly categorize users as corporate users, personal email, and institutional users such as schools and universities. Corporate users and free mail users are the two larger groups here, is there a specific group among these we are looking to target?
Is this feature something we want to launch with a suite of other features that are getting launched, are there any geographic / language markets we want to focus on?
Is this feature expected to address older habits, or do something new, something people didn’t necessarily know they needed?

From the above questions, I’m going to assume that we want to

- Build a new feature
- Focus mainly on the corporate users
- Build something that reduces some trivial back and forth that occurs in offices email chains.

Existent Use cases

In an office environment, emails are the primary mode of communication.

Users use drafts passively, i.e., they get distracted, or something else requires more urgent attention than finishing an email, and that then sits in their draft box.

Few active uses of drafts are:

Collating / Collaboration on an email by multiple team members
Sending an email to a team member for review
Longer or sensitive email that the user wants to reread before sending.

To improve drafts, I think working on the active uses of drafts can be a high-impact project.

Two of the features described above that I think can be powerful adds are:

Use Cases to explore

Email review - done by team members and managers

Email review - done by AI for tone and sentiment

Action	Solution	Value	Complexity
Manager / Teammate Review	Smart send: The ability for a teammate/manager to forward an attached draft with the original drafter’s name, avoiding the trivial ‘looks good’	Medium	Medium
Tone review	Have ML/AI decipher the tone and language of an email. Can suggest edits to appear more confident, more formal, or tone down an aggressive email.	High	High
Collaboration on email draft	Create a space for a collaborative draft where teammates can draft an email together. This reduces the overhead of formatting a final email	High	Medium

The feature I like the best is the Tone review, however, I would lean towards a collaborative email draft as the best bet.

My rationale for the choice is that it will have a broader impact on productivity and will be used more often than the other two features. Additionally, it is in line with the general collaboration-oriented culture that most companies are looking to promote.

Success Metrics:

Given this is a new feature, in addition to the primary metrics that are used to measure the health of the mail client itself, I would want to measure how well this is adopted.

Metrics that will help measure the success of this feature can be in decreasing order of priority

Fraction of emails that a group sends which are from collaborated drafts
Amount of time collaborative drafts sit in people’s drafts folders
Number of times collaborative drafts remain unsent
Percentage of users who use this feature at least a certain number of times

Other data worth tracking is the average number of collaborators on an email to give a sense of technical considerations to be made on further iterations.

Risks:

As is the case with any collaborative software, audibility is an added layer of complexity when building this feature and versioning will need to be considered.
There is a balance that needs to be achieved between how much control is given to each contributor before finally sending an email. From a user perspective, they may feel underrepresented if the email is sent without them explicitly signing off, versus from a team perspective, one user can hold back sending the email indefinitely.
Many corporates such as financial services may have proprietary software sitting on top of their email clients to prevent information leaks such as software-based Chinese walls. These scripts may unintendingly hinder a full adoption of the feature.