Yes, YOU can write docs as cool as Twilio’s

After attending Write the Docs 2018 (Portland, edition) and watching a YouTube video, I’ve got it figured out: How you (i.e. anyone) can create documentation as cool as Twilio’s–a consistent entry in lists of the best API documentation examples.

All you need to do is follow these six steps:

  1. Think, plan, and work iteratively.
  2. Infuse user research and analytics into your entire writing process.
  3. Treat your documentation as a product.
  4. Do what’s best for your documentation customer.
  5. Create and support a writing process that enables contributions from outside of the writing team.
  6. Build the CMS that supports all of the above.

That’s it!

You’ll need to understand that it took them a few years to get to where they are, today. And, it sounds like they’ll keep moving the bar.

But, don’t be discouraged. The sooner you start, the sooner those three to five years will be behind you.

A little background

Twilio’s Kat King opened Write the Docs 2018 in Portland this May, starting the conference off with a strong sense of user testing and research. That always gets my attention at a tech-writing conference. The next session by Jen Lambourne continued by describing how she applied user research on documentation. I’ve been going to tech writing conferences for quite a while and I can’t recall going to one with such a user-research focus.

I’ll make references to these videos, in what follows in case you want to go straight to the source (something I always encourage). The practices they describe, however, are straight out of user-centered-design. You can also see some of my favorite books on the topic for more background.

What’s noteworthy here is that the Twilio team has applied them successfully to their documentation–proving, if nothing else, that documentation works like any other product (and perhaps works best when treated as one).

So, here’s how the steps work:

Continue reading “Yes, YOU can write docs as cool as Twilio’s”

Agile documentation in my email

Just shortly after writing a few posts about Agile documentation, I received an invitation to this seminar in Atlanta…

Applying Agile Project Management to Information Development

I don’t know if they have some search-engine magic going on (like Amazon) or (more reasonably) it was just a coincidence in that they were announcing their seminar a couple of months before the event.

It’s great to see that Agile project management is getting some attention in the documentation realm. At the same time, it’s important to understand that Agile isn’t something you can just bring back from a seminar or learn from a blog post.

When I first jumped into an Agile development shop as a tech writer with a solid background in plan-driven project management (a.k.a. Waterfall), I talked to a few Agile coaches and consultants. They all said a 5-year time frame was a reasonable expectation to move from a plan-driven to change-driven methodology. It might happen more quickly, or it might never happen, no matter how hard you try, because requires changing how many people not only view projects, but, literally, think about the world. That was 5 years of constant coaching and encouragement, by the way.

But, in any case, it’s good to see it’s getting some attention. You can’t get to the end of a 5-year transition if you never start.

Agile documentation in the blogosphere

This is a tough topic to research. Using the words “agile” and “documentation” in the same search query returns a mixed bag of results. Some citing that the Agile Manifesto‘s goal of “Working software over comprehensive documentation” to mean that Agile products need little to no documentation. A more realistic and practical interpretation of that line is to treat documentation as any other component of the project.

A project or product should not have features that do not add value. It should not have code that doesn’t add value. Nor should it have documentation that doesn’t add value.

I’ll go along with that, but when it comes to reading what others have to say about technical writing, searching for “agile technical writing” returned this collection (in no particular order):

That’s quite a list and a range of perspectives. They range in age from 2 to 12 years old, with an average of 6.3 years.

Some early observations:

  • The earlier ones seemed to explain Agile more than the more recent articles.
  • They all offered tips for writers in one form or another
  • About half were written as first-person experience reports and the other half in more of a third-person, instructional format.
  • Most were written in a narrative format with a few as just bullets or bullets with some details.
  • Only a few cited additional references (although most had embedded links to related topics)

I’m not sure what conclusions to draw from this, yet. For perspective, I wanted to do a similar survey of academic literature, but searching for “Agile technical writing” didn’t produce much to read. Hmmmm….

Agile documentation in the real world

In Agile documentation in practice I described what worked for me, which works, when it works, but it’s not without its challenges. However, if you can get these right, the process should be smoother.


First and foremost, it works best when everyone is on the same page. When stakeholders have different ideas of quality, content, scheduling, iterations, or priority, things become difficult. This should come as no surprise, but it’s worth repeating because without a common understanding of the ground rules by all the stakeholders, those disagreements complicate all of the other areas.


In Agile documentation in practice, I mentioned how the traditional writing process was not what worked for me in an Agile shop. When I matched my writing process to the coding process as much as possible, things became easier. Once I was able to plan and describe my process in terms that made sense to agile software developers—something that I couldn’t do until I embraced the process for myself—it was easier for them to understand what I was doing (and when I was doing it).  Reporting progress clearly and frequently is an essential part of a successful process (Agile or otherwise).


A big part of the process consists of the tools used to execute the process. Agile is designed for changes. Changes in requirements. Changes in features (content, for documentation). Changes in audience. To name a few. The tools used to manage the workload and the content must accommodate the changes you want to accommodate as easily as possible.

Which tools are best? That depends on your process (and is worthy of a topic or two in itself). Ideally, the process will define the tools, not the other way around. In reality, however, this is often difficult to apply in practice. At least, with a clear process, you’ll know what your tools will need to do.


Collecting and analyzing data must be integral to the process. Data such as audience requirements, product road map, market climate, at the high-level and direct customer feedback, usage and satisfaction metrics, and production metrics at the detailed end, to name a few. Without a constant calibration, the writing process becomes increasingly disconnected from reality—and remember, reality is a moving target, which is why you need to keep your data up-to-date.


If you’re an Agile shop, this shouldn’t be a problem you can’t handle because the sprint planning process takes this into account. Each sprint has its velocity (capacity) for development tasks and it will also have its velocity for technical writing. This always begs the question:

What is the right ratio of developers to writers?

It depends, of course. But if you have all of the above, you might not have enough writers to produce everything you want to deliver, but you’ll have the process to do the best you can with what you have and the data to know if you need more or fewer.

Agile documentation in practice

The last  couple of posts seem to state the obvious. They are , basically, variations of know your reader/customer. So, how hard could it be?

In my experience, there’s never time to do all of what is suggested. As a professional writer, the “topics-to-write” list has always been longer than the time available to write them—often by an order of magnitude.

Exactly the situation Agile was designed for.

I would think that writers would flock to Agile, if it’ll solve the problem of “so many docs, yet so little time.” And, they might, if given the chance.

Yet, in an unscientific study of documentation planning methodologies, they all look like they are based on a plan-driven (a.k.a. waterfall) project management model:

Plan, Write, Revise, Edit, Publish

…as though publish was something final. In Agile documentation, that might describe the process to publish a single topic, but it doesn’t scale up. If we call this sequence write-a-topic, in the larger scale, the documentation process looks more like:

Plan, Write-a-topic, Review, Repeat

Stopping every so often to step back and look at the view (i.e. take care of more strategic planning).

What works for me

I try to organize my documentation as much like the software development process as possible, but it took some de-programming to break the waterfall method into which I’d been indoctrinated.

It works best when I can define my documentation projects as epics, sprints, stories.

An epic is a large-scale project or functionality that spans multiple sprints and consists of many stories. For example, produce the documentation for a new feature. While the developers design the feature, I’m working with them to design the documentation—keeping the Agile methodology in mind. I  won’t be able to write it all at once, so I work with the product owner and the development team to figure out what functionality will be available when (in sequence, if not by date), what topics the customer will need and in what order (they can’t read it all at once anymore than I can write it all at once).

With that information, I define the stories and prioritize them with the product owner such that if I have to stop at any point along the way (e.g. to start on the next project), I’ll know that the most valuable topics have been written and the rest can wait (or they really weren’t as valuable or important as I thought they were).

The sprint planning and reviews keep the priorities in focus, but that often means that the epics are not done linearly. One documentation epic was about 2 months of work, but took 4-5 months to complete because of intervening priorities. Fortunately, I designed the content such that it was always publishable (shippable, in Agile terms) and the overall content set just kept getting better as I wrote it. Eventually, it was finished—as was all the intervening, higher-value content.

As Hannibal of the A-Team would say, “I love it when a plan comes together.”

Agile documentation customers

I continue the reflection I started in my last post, reviewing Scott Ambler’s article from 2002, Can Documentation Be Agile?

To no surprise, the customer emerged as being a critical element in Agile documentation. The points in Scott’s article refer to the Customer as though the customer is singular. Users, the user, etc. are also used similarly. In most cases, however, (i.e. with any luck) there is rarely only one customer. So, how can we understand all those people?

Find the center of mass

In physics, large bodies are modeled around a center of gravity or center of mass. This notion facilitates the study of very large objects, planet- or galaxy-sized objects in astrophysics, by modeling them as a single point with the mass of the entire object.

As with planets and galaxies, customers can also be modeled in a similar fashion. Many customers can be (and are frequently) modeled as a few personas or market segments. This is common in fields that deal with large numbers of people, such as marketing and advertising. In these cases populations are described by what they have in common and treated as a single persona. Does that mean every person in a segment’s population is the same? Of course not. But a well designed persona or segment describes the common attributes of the group sufficiently for its intended use. Just like knowing the location of an object’s center of mass will let you understand its properties and behaviors when acted upon by external forces.

One drawback to this method, is that when you miscalculate an object’s center of mass, it won’t behave as you expect it should. Its behavior will be consistent with its actual center of mass (whether you know where that is, or not). Where I’ve seen personas and segments go awry in technical writing is, inevitably, when they were not applied or not defined well.

Base personas on research

Just as an object’s center of mass is found through measurement, a population’s or segment’s properties are also determined through measurements and analysis (a.k.a. research).

If you want to treat a population as a single entity, you must measure it precisely.

But measuring customers is not easy. So what’s a writer to do?


Successive approximation is a method for improving and refining a measurement and is an important part of Agile. Measure, test, observe, refine, repeat (or pick your favorite method for this). With this approach, you can continuously improve your customer model such that you can treat the customer as a single entity. There are several ways to accomplish successive approximation methods, but the key factor is that you assume that you must evaluate and iterate and that these aspects are baked into every process.

So, one of the first steps toward having agile documentation is having the ability to iterate (cheaply and quickly) and the ability to measure and understand your customers.

But, we all know that. Now, to figure out how to make that part of the process.

Agile documentation practices

Scott Ambler started an interesting conversation about agile documentation on LinkedIn  a short while ago. I visited Scott’s site, to see what else he had to say on the subject and I came across his article from 2002, Can Documentation Be Agile?

In his article, he lists 12 points that can help documentation be agile.   Scott’s article outlines a view of writing that I have not seen in practice in the 14 years since he wrote it. In my careers as a developer and a writer, I’ve seen bits and pieces of his list in practice to see how they might work together. I’ve also seen many obstacles that keep them from becoming more widespread. So in that context, I reflected on his points to imagine what such a world might look like.

I think it starts with us, the Writers, who need to look at the job from a new perspective.

Writers are more than “the scribes”

First, writers are not just “wordsmiths.”  While we take pride in our ability to craft great content, that’s not where we add value. Our value comes from what we craft, not simply that we craft. Likewise, software developers don’t add value just because they code, they add value because of what they code.

So, how do we know what to craft?

Focus on the customer

His first point is where writing really starts and ends—the writer’s audience and customer. Writers need to know their audience first hand. They need to walk in their shoes. More importantly, they need to make (or be granted) the time and resources to do this. Every writer? Yes, I think so, at least to some degree. Why? The longer a writer goes without walking in the customer’s shoes, the more assumptions creep in and take hold. It starts off slowly, but eventually, the writer’s view of the customer gets out of sync with the real customer. Markets change. Customers change, Products change. It’s a full time job keeping them all in sync, but it’s a necessary one.

Keep it simple, but not too simple

Of course, where that fine line falls is the source of constant debate. But, why is that such a debatable topic? I refer back to the previous point. If you know your customers, you know what they want. If you don’t, you make assumptions, and so the discussion becomes one of assumptions and not data. Hence, the contention.

Another aspect of this which provides a challenge to a writer is when one topic needs a lot of content to be “simple enough,” while another requires very little. The former looks lavish and the latter seems incomplete—until you use the, of course. But I’ve seen this unevenness be the topic of contentious discussions. Why? Because, audience data to the contrary was lacking.

All the more reason for writers to get out and meet their customers.