OverviewUser interviews and researchFrom daily message to straight actionsEnhance production lifecycle featuresimprovement in team productivity
Case study #2

Improving development productivity by enriching processes and communication within the team

2022 Zigi.Ai Ltd.

Overview

This case study is about improving development productivity; the product reduces cycle time by ~65% by enhancing processes and communication within the team.

I started by interviewing and researching the workflows in different development teams. Based on their needs, I improved the daily message for developers and prototyped Slack-based tools such as live notifications for PR review, connecting code to tasks, etc. These tools went into development and are continuously improved based on user data and feedback.

Development team’s need for cross-tool communication

After talking with the development team, I understood different workflows. Every developer/team works differently; each workflow has different requirements. The core of all ideas was to lower what developers need to talk about - like asking for a PR review, notifying when they finish a task, etc.

Users and audience

  • Developers
  • Developer team leaders

Roles and responsibilities

  • Product design - Myself
  • Founder and CEO - Nir Benson
  • Co-founder and CTO - Omri Rotem
  • VP Marketing - Zivit Katz Benson
  • Content consultant - Naama Oren

Unique factors

  • Limited to the Slack Builder Kit features
  • No option to record (Slack-based product)

Scope

  • Not replacing or mirroring code management and workflow tools, rather improving their productivity
  • Stay in the Slack tools - no web interface

The process

  • Started by interviewing developers from different companies and with different team sizes, understanding their workflow pains and getting current product reviews.
  • Improved the current product while  up with additional features.
  • Testing: getting user reviews, improving, implementing, and testing what works.
  • Improving current version and re-implementing, while continuing with constant user interviews (step 1).

User interviews and research

When exploring the current daily message, I understood its main issues - too much information, unclear information such as actions, timeframe, events, kind of items, etc. I moved from long awkward message to short, clear quick actions in two sections.

Initial version

When I joined the team, the main product was the daily message for developers:

General research questions:

  • Describe your morning routine and meetings, duration, what works well, and what doesn’t.
  • Describe what interests you when you start your day.
  • Describe your morning report to your manager and team. Are there any formats, tools you use, etc?
  • Do you use any bots for Slack? Please describe them, what they do, and indicate if they work well.

Current product research questions:

  • Look at these "daily reports" and share your thoughts.
  • What do you feel about each bullet? Is it relevant?
  • What do you think about each action? Is it relevant?
  • What would you improve?
  • Please look at these "updates since yesterday" and share your thoughts.
  • Share your thoughts about these "sprint tasks". 

From daily message to straight actions

When exploring the current daily message, I understood its main issues - too much information, unclear information such as actions, timeframe, events, types of items, etc. I moved from long awkward messages to short, clear quick actions in two sections.

Improved version

After several versions, each tested and reviewed by users, we came up with a new version:

Issues and their solutions

1. The developers felt there was too much information and it was all mixed

  • Instead of a morning message with many details like sprint number or even a "good morning" message, the new version has two primary actions.
  • I formulated each phrase with bold text and a precise action.
  • I limited the number of items in the message, and moved the rest to a "view more" section, accessible via a button.

2. Developers didn't understand when each event happened and what to prioritize; they wanted to understand which action was most important and why

  • In each bullet, I added  the time and actions arranged from newest to oldest: a bold section that gives information about when it happened or how much time has passed so that the developer knows how long it's been waiting.
  • The decision to focus on two primary sections limited the message and removed non-relevant information and actions

3. Developers didn't understand what related to Jira and what to GitHub; what from this list were tasks and what were PRs

  • The research emphasizes the essential guideline of mentioning the item type with "PR" or "ticket" when giving information to the developer.
  • Eventually, this message changed to only PR-related actions, but the need to mention the item type was emphasized here, and in  other "Zigi" features.

Enhance production lifecycle features

The second stage of the work was to produce tools for automatic actions that would replace information spoken or written and give the user essential information to enhance the development process.

Team leader actions

To prevent team leaders needing to ask or the problem of indirectly discovering delays, we created messages that reach the team leader and update them on tasks or code reviews that are stuck in one of their team members’ workflow.

PR lifecycle

Developers use “pull requests” in day-to-day work to initiate integrating new code changes into the main project repository. During this PR life cycle, they ask for a review and approval from teammates, and review and approve others' code. Each team/devs asks for this differently, via talking, writing messages, using code management, etc.

These "Zigi" messages help "move the ball from one to the other" in the PR flow - it's live messages that explainPR needs, such as "PR needs review", "PR commented", "PR merged", etc.

Improvement in team productivity

The implementation of Zigi into companies resulted in teams/devs who used the product having significant improvement in their productivity.

Active users

Zigi implemented the product with 280 users in 70 companies (as of Nov 2020). In these companies, the daily use of various features was 65%.

Cycle time improvement

We examined usage for Zigi’s users who performed the most actions with the product over two months, compared with their previous two months not using Zigi. Here is what we found regarding the reduction in PR cycle time. In the users who used Zigi, the average improvement was 75% reduction in cycle time.