Home

Saving time with stored knowledge

Subject matter experts got a productivity boost when answers became easily available to all

For more than a decade, Skai’s solutions have powered data-driven insights and marketing execution for some of the world’s most sophisticated consumer brands. Skai’s platform includes a suite of AI-enabled products for market intelligence, omnichannel media activation, testing, and measurement that enable teams to accurately plan strategies and benefit from connected omnichannel launches. Formerly known as Kenshoo, the company rebranded after its acquisition of Signal Analytics. With its expanded product suite, Skai is unifying data and helping companies better understand their consumers in real-time. It has seven international locations and is backed by Sequoia Capital, Arts Alliance, Tenaya Capital, Bain Capital Ventures, Pitango, and Qumra Capital. All told, the company has over 300 engineers spread across eight sites, three countries, and two continents.

Screenshot of skai.io

The problem

After 15 years of building, two acquisitions, and a rebrand, Skai faced a knowledge challenge. The number of products and projects Skai’s teams were working on had started to proliferate. 

“Our teams work collaboratively on some older, shared, monolithic codebases as well as many newer, smaller, team-owned projects. This amounts to a lot of internal knowledge that has to be shared across teams,” said Tzach Zohar, a systems architect at Skai. “Newer teams spend a lot of time trying to figure things out, and veteran teams find themselves repeating the same answers or chasing outdated documents.” 

From common error messages, to naming conventions, to legacy processes, certain common questions arose over and over again as new teams tried to learn the existing system and deploy fresh updates and ideas. 

Skai realized this challenge was an opportunity for improvement and innovation. It needed a solution to help them document institutional knowledge, share knowledge across teams, and find answers faster.

The approach

Skai didn’t come to Stack Overflow for Teams right away. At first, it tried Google Sites, GitHub wikis, and several other documentation solutions. It saw partial success and gradual improvements here and there. But it continued to struggle with specific recurring technical questions related to writing code or troubleshooting a system. 

Skai’s developers went looking for “institutional knowledge” like this by informally asking someone on the team. However, remote working culture made this method challenging, especially for new employees. 

Dedicated Q&A channels in Slack offered a partial fix, but suffered from poor searchability, low signal-to-noise ratio, and poor visibility. Developers felt helpless, and subject matter experts (SMEs) answered the same series of frustrated DMs over and over again. Or worse, they wasted time searching through random messages and documents with the vague feeling they answered something, but not quite remembering when and where. 

Faced with this dilemma, Zohar and his team set out to map the key pain points and sketch out a system that would meet their needs and provide a sustainable benefit to the company. They came up with a list of seven attributes an ideal solution would offer:

Centralized

Where developers could enter queries and receive responses at an appropriately technical level

Protected

A single entry point to search for all tech-related questions

Q&A formatted

All knowledge organized as questions and answers (as opposed to Wikis / Guides / Channels)

Searchable

Tags / keywords can lead me to existing answers

Accessible

Everyone has read and write access

Moderatable

Includes mechanisms to ensure quality (delete bad content, improve readability etc.)

Updatable

The cost of adding / updating answers is low; low barrier for contributing content

Code-first

Good code formatting

With these requirements in mind, one familiar candidate seemed like a good fit: Stack Overflow for Teams.

Skai had two initial reservations: 

  1. Developers are used to referencing Stack Overflow, but would they actually ask and answer questions in an instance of Stack Overflow for Teams? 
  2. Would a 300-person organization generate a critical mass of knowledge necessary to make Stack Overflow for Teams the go-to destination for the team’s developers seeking answers?

The results

Skai planned and executed a 60-day proof-of-value test to understand whether Stack Overflow for Teams would work for it. 

Stack Overflow helped Skai to identify some key topics that might produce a high volume of activity. They also identified some key SMEs to jumpstart the activity on the platform by answering some questions. 

15 subject matter experts (SMEs) took to Stack Overflow for Teams, each with a mandate to answer the top 3-5 questions they’re most frequently asked. Within a week, Skai had more than 50 Q&As, validated and labeled with the relevant tags.

Technical teams typically face an adoption hurdle. But because Stack Overflow is already second-nature for developers, the transition to Stack Overflow for Teams was easy.

Tzach Zohar, systems architect at Skai

Adoption quickly extended to the engineering team, who began searching Stack Overflow for Teams instead of Slack regularly.

To help move those questions off of Slack, Stack Overflow for Teams includes a Slack integration that allows users to respond to questions on Slack with an automated entry in Stack Overflow for Teams. The team also responded well to gamification encouraging participation, with prizes awarded to the most active users on the platform. This created positive buzz and gave recognition to early adopters. 

An enduring benefit of the platform was a boost in knowledge reuse. Stack Overflow for Teams has a metric, Knowledge Reuse, that measures the people who come to the platform but don’t ask a question. They are able to leverage existing information from the knowledge base to solve their problem and get back to work. Skai is seeing this happen 42 times a day. That’s 42 times a day employees don’t have to interrupt one another to find a solution and make progress on their tasks.

By the end of the 60-day trial, Skai saw additional promising results:

More than 350 questions posted and answered
50% of the posts were organic, i.e. not posted by SMEs but actual questions and answers that would unblock a developer in need
Question answered, on average, in less than 1.5 hours
About 1 in 5 developers at the company visit each day
More than 200 daily searches

In particular, the amount of knowledge reuse stood out as a healthy sign of a meaningful behavioral change. Engineers visit Stack Overflow for Teams, search for answers, and remove obstacles to their work. This habit prevents them from disrupting someone else’s workflow, or sending pings after-hours.  

SMEs, who had put in some extra effort up front, reported noticeable time savings, especially when a “hot topic” had a thorough answer on Stack Overflow for Teams.  

Today the knowledge reuse, as measured by number of searches and unique daily visitors, are still on the rise for Skai’s engineering organization.

Metrics from Skai’s Stack Overflow for Teams dashboard

The plan going forward is to continue working on backfilling older knowledge and growing the habit of searching on Stack Overflow for Teams until it’s second nature. 

The team is also leveraging features like Articles, which allows them to move beyond just Q&A to include long-form content like how-to guides, deep dives and manuals, and research in its knowledge base. Skai is collecting long-form content, which had previously been scattered across Google Drives and GitHub repositories, and putting it in one central, searchable location on Stack Overflow for Teams. 

Skai has plans to make Stack Overflow for Teams available to product, client success, solutions, and even the sales team. As Zohar put it, “As usage of Stack Overflow for Teams continues to grow, it continues to prove its value.” 

To find out more about Skai, visit: skai.io.

What to read next

Sign up to receive helpful content and resources from Stack Overflow