What I Learned Building Knowledge Bases at Microsoft

“Knowledge is life with wings.” — William Blake

I learned a lot from building world-class knowledge bases at Microsoft over 25 years.   Mostly I learned how to learn faster by organizing incredible libraries of profound knowledge.

But also I’ve learned that knowledge is an incredible competitive advantage.

The value of your skills is the value of the problems you solve.

One of my most amazing experiences at Microsoft was working on the Microsoft knowledge base.

It was a library of many thousands of articles written by smart people who wanted to solve customer problems.

Imagine a library of thousands of articles created by some of the most amazing experts.   What an incredible way to share and scale expertise.

I remember how powerful it was to learn that I could write an article that could potentially solve issues for many thousands of customers.

The big surprise to me was how profound knowledge could be easily put into a self-service library available to the world.

My Unique Position

Normally, one lead would only work on one knowledge base.

I had the unique position of being in charge of multiple knowledge bases, and then changing teams early, where I had to then whip new knowledge bases into shape.

This experience helped me learn how to manage vast seas of knowledge faster.

Imagine hundreds or thousands of articles that directly impact your business.

The stakes were high because the quality of the knowledge base directly determined how many customer could solve their own problems with the help of our library.

And it was not just a matter of having great articles available.

It was a matter of how easily users could find the right article for their problem or tasks.

I really focused on things like naming conventions, creating browsable pages by topics and subtopics, and using feedback for continuous improvement.

Relevancy is the high order bit—if an article is not relevant, then it doesn’t matter.

Speed of Insight

I focused on speed of insight.   A lot of the best knowledge we had was stuck in the heads of experts that had years of experience.

The challenge was that some experts weren’t really writers.

And some experts were perfectionists.

And some experts wanted to only write long articles that were difficult to edit and a nightmare to maintain.

While some of the long articles were true works of art, we couldn’t really scale that approach very well, in any sort of consistent way.

Plus, it intimidated new writers, when they thought they had to write epic articles to solve a problem.

Gradually, I convinced teams that we could write smaller, more focused articles and use feedback to improve them more often.

In some ways, this was an early experience of Agile content.

I focused on speed, usefulness, and quality, and I found that I could improve quality by shipping faster and reviewing articles more often.

So each knowledge base ultimately would become a living library of knowledge and know how.

Knowledge Nuggets (Modular, Reusable Content)

This modular approach to content was very empowering.

Suddenly, articles didn’t feel heavy anymore.

And because each article was focused on solving a problem or teaching an idea, it was very easy to measure how effective it was.

It was no longer about what an amazing writer the expert was.   II t was about how well did we share and scale an incredibly deep piece of knowledge.

In fact, many of the experts were not great writers.   But the surprise is they were great teachers.

And they were great leaners.

In fact, I never met a true expert that was not first and foremost a great leaner.  Know-it-alls fade over time.  Learn-it-alls are forever.

So creating a world-class knowledge base became a game of pairing the right people up to unlock their greatness.

Looking back, it would have been great to pair up experts with a strong writing team, but I can also say there is something very special about articles written by the expert vs. articles written for, or even with the expert.

What made this even possible though, was choosing to focus on small, modular, reusable knowledge nuggets that could easily be managed and maintained.

It wasn’t easy but once everybody realized that their 5th version of the article would be much better than their first, it helped a lot to break the bottleneck of our content and flow more articles to the Web.

I made it real for people by showing that we actually would update our articles and even make it fun.

Action vs. Reference

The first thing that really struck me about the Knowledge Base is that all of the knowledge could basically be organized into two different categories:

1. Action
2. Reference

This was a very important distinction because some articles that tried to do both, were pretty ineffective.  I found it much better to have the conceptual and overview articles just focus on explaining the topic in depth and detail.  And then a separate set of articles would be focused on action, such as a set of steps, like in a How To.

Before this distinction, articles were a real mess.

It was very difficult to read an article that was mixing in steps that you weren’t ready to perform.  And it was difficult trying to perform steps where all of a sudden there was a long explanation about a concept.

I learned that it was easier to learn faster by separating action from reference.

In addition to the knowledge bases I was working on, I started organizing all of my personal information into Reference vs. Action. Sometimes there was overlap, but mostly I tried to create very clean action steps separate from information that was more for reference, which could be things like cheat sheets, checklists, stories, overviews, concepts, etc.

I carried over this concept of reference vs. action into all sorts of areas in my work and life.  In fact, it’s one of the ideas I baked into my productivity system, Agile Results to help people shift from “think mode” to “do mode” and take action easier, while having the reference information at their fingertips.

Strategic Content “Types”

One of the early things I noticed was that knowledge base articles gave the reader a clue about the type of article within the title:

• INFO – What is the X Used For
• HOW TO – Connect to the Database
• PRB – Why Does X Show an Error Message (PRB stood for Problem)
• Solution – Impersonating the Original Caller
• BUG – The X Mishandles the Call to Y
• Checklist – What To Check For When Doing X
• etc…

It was only a handful of article types.

But that along instantly  helped to group a set of articles into whether the article was going to give me an overview of the concept or show me how to perform a set of steps to solve a problem.

I realized these were not just any article “types”.   They were strategic types.  These were specific article types that had high value.

You can imagine the same thing with Google and how Google’s search engine has collections of strategic types of content:

• Videos
• Images
• Maps
• News
• Books
• etc…

I was part of an early pilot effort at Microsoft to build a library of How Tos.

What surprised me was how with the right library of articles, written by the right experts, our team could change the world, by sharing and scaling deep expertise.

It’s actually very difficult to build a great set of action steps for a human.  But when you do this well, it’s one of the most amazing ways to share and scale knowledge.

Collections of Content Types

I was able to actually measure our customer satisfaction and success based on how well our knowledge base performed.

When customers would easily browse a collection of strategic content types (How Tos, INFO, etc.), customers were able to easily solve their problems and learn about new technologies.

Empathy with User Scenarios and Real-World Problems

The big factor here that really made a difference was using experts to write, the same experts that were working with customers through real problems.

Everybody that wrote articles had deep empathy for customer pain.

You can’t buy that.  You can’t outsource that.

What we did was lean down the writing process, pair people up, and actually treat writing content like a first-class citizen.

What helped is when I was able to show the business case how content was scaling expertise, increasing customer satisfaction, reducing call times, and solving problems that used to take people days (weeks, months).

Better Product Feedback

Imagine how much better it was to give product feedback by being able to point to the most used articles that solved the worst problems or the most common problems.

The product teams could see a set of steps that customers had to take to solve problems or workaround issues.

This is way better than saying “Product X sucks” or “Feature Y sucks”   We’re talking a super easy way to see how customers are using (and misusing) the products and platform.

In a way, each article served as a lens into customer behavior, because we updated articles to reflect what we learned with customers.

You Can Learn and Evolve Products Faster with Content

I used rapid content to solve customer problems around getting started, using a particular technology, and solving complex problems.

This is exactly how I helped bake future solutions into the product.

And I don’t mean just simple things.  I mean complex things such as how to solve complex security issues.  In fact, our teams ended up feeding heavily into the future of ASP.NET by providing steps we were using with customers.

A great light bulb in my mind lit up the day I realized that content could be used to explore, experiment, and test future solutions that could become product solutions.

Reducing Burnout of Our Best Experts

The surprise is how the worst, most complex problems are suddenly easy when the expert shows you how they solve it (again, and again, and again).

This was another big advantage I need to call out.

By building a smart library of articles, the experts didn’t have to keep solving the same thing over and over.

Some say the definition of burn out is solving the same problems over and over.

Before the articles, the experts would have to walk customers through repetitive solutions.  And not just customers, the experts would also have to walk their peers through solutions.

By writing articles that solve the repetitive problems, we freed our experts up to focus on higher order problems and keep learning and growing.

Faster Ramp Up for New Talent

A well-written article had the extreme advantage that somebody new to the team could quickly learn key concepts and steps for even complex technology.

I worked across a lot of technologies, so one of the best ways I found to work across “domains” or areas of knowledge was to read through another team’s knowledge base.

I loved the fact that with a well-built knowledge base of profound knowledge, as people come and go, the knowledge base could actually grow with new insights and fresh perspectives.

At the same time, we didn’t keep losing the amazing knowledge that so many experts had built up over time.

Nothing was worse than asking a question for a profound problem, only to have somebody say, oh, you need to go ask so-and-so about that.

On the flip side, nothing was better than finding that articles that spoke to your exact question and nailed it.

Thank you so-and-so for sharing your deep knowledge and wisdom in the form of an article.

It reminds me why reading and writing are actually valuable skills.

Questions, Tasks, and Cornerstone Concepts

One day, Microsoft Human Resources reached out to interview me.   They wanted to learn how I learned fast.

I remember going to the whiteboard to map out my process, and I realized it all came down to the idea that you can learn anything faster by focusing on:

1. Questions
2. Tasks
3. Concepts

Want to learn something better?  Ask better questions.   Want to turn that information into useful knowledge?  Apply it against challenging tasks.

Want to connect the dots like a constellation and light up your knowledge nodes  Learn the key concepts.

One of the most important things I learned is how the world’s knowledge for any domain could be organized by question, task, or concept.

This really surprised me, but it was something I figured out during many “special projects” where I was asked to organize massive quantities of articles into useful libraries.

I found it was possible for lesser products to beat better products simply by building better knowledge bases that addressed user’s questions, tasks, and cornerstone concepts.

Here’s the thing—you don’t need to cover everything with content.

I learned that I need to focus on smart mental scaffolding to chop a domain down into size so that the community could advance the space.

This is where I got to really practice the 80/20 rule.   I focused on seeding with articles that addressed the most important questions, tasks, and concepts with the idea that the community would build out from there.

And this turned out to be true.    The key here was to first and foremost make sure that it was a strong knowledge base to start with.

Getting Started, Core, Common, and Niche/Advance

This is something I learned over time simply by grouping hundreds, thousands and tens of thousands of articles into categories and buckets and trying to find them fast.

I also learned this by doing super competitive assessments where my job was to help our products and technologies win in that space.

I learned that by grouping content into a simple framework, I could easily help any user through any complex or challenging space.

The framework I used is:

• Getting Started – This is where the content to help a user dive in and jumpstart their learning journey, by both learning key concepts and learning by doing.
• Core – This category is about figuring out the most fundamental knowledge that somebody needed to be successful in this space and to start off their learning journey from solid ground.  Really this is a well organized set of fundamentals.
• Common – This is where the 80/20 rule kicked in.   I would first build an ontology or simple set of topics and sub-topics that effectively serve as an outline of the domain or space.   I would base this on all the preexisting knowledge to learn what the Hot Spots were in terms of the most important topics.   The beauty is that for each of these topics, I could focus on principles, patterns, and practices to effectively build a learning framework to advance the space.  The goal was never to cover it all. The purpose was not to be the driver’s manual, but instead to be the driver’s guide.   You need both the driver’s manual and the driver’s guide to truly master any knowledge domain, but I focused on building better driver’s guides because that’s almost always where the gap is.
• Niche / Advanced – This is about covering the advanced topics that people don’t always need, but when they run into them, they are incredibly thankful that somebody has laid down the path and actually written about it.   II t was important to carve this out from the core, because it was very easy for experts to suddenly write about all the advanced topics they know and skip over getting started or the core stuff or even the common challenges in the space.

The big surprise is when I did audits of content for big topics like security across Microsoft as well as other companies.   Early on, I learned that a company like Apple was good at Getting Started, and even some of the core stuff, but you could easily then find big gaps.

On the flip side, Microsoft was very good at covering the more advanced, but missed a lot of the getting started or the common use cases.

As a result, I focused a lot of my efforts on shining the spotlight on the gaps and opportunities so we could make it much easier to support a customer’s learning journey from getting started through core knowledge to common scenarios and even advanced use cases.

In the end, this effort showed up in things like competitive studies as well as in terms of customer satisfaction and confidence in the platform.

That’s powerful stuff.

Call to Action

1. Practice the habit of organizing your information in terms of action and reference.
2. Practice organizing your information in modular, reusable knowledge nuggets.
3. Practice learning anything faster by mapping out the key questions, high value tasks, and the deepest concepts.  Tip – you can build better maps by checking with experts, and real experts love to share what they know.

You Might Also Like

10 Best Project Management Books for Real-World Results
Are You Used to Delivering Daily and Responding to Change?
How Prescriptive Guidance Helps Win Competitive Assessments
How To Brainstorm Better with Information Models
What I Learned About Agile Project Management at Microsoft