Dogfooding, or how humbling it is to use your own product for a month

Source: Unsplash.com. Photographer: Jairo Alzate.

I never much liked the phrase “eat your own dog food.”

It’s a popular phrase (and a great idea) in the startup world, first made known they say by Microsoft but made known to me by the always-smart Joel Spolsky sometime after 2001, when he explained it like so:

Eating your own dog food is the quaint name that we in the computer industry give to the process of actually using your own product.

At Airstory, we decided to make June 2017 the month in which we’d do some serious dogfooding. The plan: the five F/T members of our team would each write one blog post a week for four weeks. We’d put ourselves in the shoes of one of our primary segments of users (i.e., pro bloggers) to see if and how we were — or were not — providing value to them. And we’d compare notes at the end.

We’re at the end now.

We’ve put all of our reactions, ideas and thoughts into Airstory as cards. I decided to make this the subject of my final post of the month. I thought I was getting off easy — suckers slogging away writing original posts while I turn their cards into my post with the greatest of ease.

Joke’s on me. Turns out this is the toughest post I’ve written all month.

Because it’s humbling AF.

Lemme just say there’s a reason businesses don’t do this. You’ll see why.

Biggest takeaway: OMG dogfooding is a good exercise — as long as you document your reactions along the way.

Most unexpected takeaway: While blogging, my cofounder Steven — who built 99% of Airstory — forgot about some of the features he’d built. In addition to forgetting to start in our template marketplace to find a blog post template (to battle the blank page), he said:

I remember complaining about how it would be nice to only have a view of cards if that is what I am working with…completely forgetting about the full screen button in the drawer (surprisingly, this little feature greatly improved my process).

Some of us have published hundreds of blog posts. Others among us had never published one. Only Lance and I had written posts in Airstory before this exercise. Here’s what we found.

The #1 area of opportunity: bloggers need a lot of help with the writing process

Of all the reactions our team had to blogging in Airstory this month, the reactions having to do with process challenges were most plentiful. Every team member had at least one note on the subject, and grand total we had 14 notes — more notes than for any other subject.

What’s interesting about this is that Airstory currently doesn’t have anything built in re: process.

So the Airstory writing process isn’t broken. It’s just non-existent. We can work with that — we can do cool things with that.

By far, the biggest challenges with writing process were 1) ideation and 2) staying on task. Steven described these problems well…

“I discovered how difficult it was to create original content. When trying to think through an idea, it was very easy to be distracted. It was like my mind WANTED to be distracted. If I was adding cards to the card library and I could see the editor (the default setup when you enter a project) my mind would be drawn to the editor.

“That’s when I started to dedicate my screen to what I was working on at the time. If I was working in the card library, I found that just by simply clicking the expand icon at the top of the library:

“I was able focus a lot more on just entering cards. That became what I was doing and my mind no longer found distractions.

“Similarly, if I was working in the editor, cleaning up my draft after everything came together, I would close down the card library, collapse the top of the page and expand the toolbar so that my workplace looked like this:

“At this stage, I no longer needed cards, so closing the card drawer and giving the editor as much room as possible allowed me to concentrate on the task at hand.”

Other challenges we found with the process of writing:

• Research overload! Described by Sarah as “getting sucked into the vortex of information,” it’s hard for a content creator to stay on topic when you start reading through so much stuff online and clipping as you go. Feature idea: A timer that keeps you on task.
• The lack of structure can be intimidating. Andrew noted that writing an essay in school felt easier because you had a rigid structure, a list of topics and a minimum word count. When trying to write start-to-finish in Airstory, he struggled; when he switched to thinking about writing a post as a modular experience — with building blocks — writing got easier.
• When your writing will be exposed / available to everyone as a published blog post, self-doubt kicks in for the less-experienced blogger.
• We need to do more than set deadlines and word count goals in Airstory. Notifications could help, but so could writing prompts and helpful reminders spread out over the course of the period between project-start and deadline.
• Tangent to-dos and sudden appearances of new ideas are not easily managed with Airstory. An idea hits you, and you have to go through a handful of steps just to get it down. Naturally, you’ll stick with your old way in this case ‘cos it’s faster.

Changing well-formed writing processes is hard — even when the old way is ultimately inefficient. In the moments when you have to choose between a) changing and meeting a new devil or b) staying the same and keeping company with the devil you know, you’re more likely to choose B. As Steven has said countless times, Airstory needs to be 10x to 100x better than the current way, or folks won’t switch.

Thankfully, we had some really cool wins in this area, too. Lance discovered and loved this blogging process:

1. Creating cards using the Researcher
2. Dragging the cards onto the page
3. Leaving cards unmerged, and then shuffling them around into an organized state, and
4. Merging cards 1 by 1 and editing the content.

And my very-similar writing process helped me start and finish a 4300-word blog post in a single day — something I’d never, ever done before. (It usually takes me months to move from idea to completion.) The post was well-shared, drawing some 100K readers. Someone submitted it to Hacker News, where it got tons of comments and made it almost to the top:

If we can help people improve their writing process — such that they can come up with ideas they’re confident are worth the time and then invest the right amount of time at each stage in the writing process — Airstory will be vastly superior to any writing platform on the planet.

Some of the ideas for optimization and/or new Airstory features that came out of this exercise:

• Turn on “Airstory Suggestions” to get a list (in some form) of places to start researching a topic.
• Invite someone on your team to give you a quick gut-check on whether a topic is worth the effort or not.
• Use data (about either Airstory users or all published content) to say, in Expedia form, “In the last 30 days, 22,447 pieces of content have been published online featuring content about {keyword phrase identified in project title and/or summary}.” This could, of course, harm or help getting started. Maybe then it’s more of a uniqueness barometer vs the-world-wants-this barometer.
• Build a side-project Airstory Idea Generator where, when you land on a good idea, you can immediately create an Airstory project out of it.
• Develop prompts to help keep the writer in the heat of the original idea — anything to keep them from being distracted, which gives the idea time to cool and start to look lacklustre.
• The Missed Deadline Punisher — once you set a deadline, you can’t change it and you face X consequence for missing it.

The core of our product (Editor + Cards) is in pretty solid shape

Before going into Dogfooding Month, we knew a few things to be necessary enhancements of the Airstory core:

• We need more robust search for our Card Library
• We desperately need to ‘mute’ or hide used cards instead of insisting the user archive as she goes
• We need to make it possible to edit an unmerged card within the document

Now we have further validation that those points are top priorities.

We also discovered some interesting stuff about how we use cards:

• Card titles are more important than we’ve given them credit for; they’re hidden when a card is dragged onto the page, which can cause challenges.
• Card content is more important than we’ve given it credit for; it’s hidden when a card is dragged into the outline.
• We didn’t reuse cards we’d created for blog posts, so the “single source of truth” that the Card Library is doesn’t apply for bloggers as it will for marketing, product, legal and HR teams.
• It’s rather painful to move cards between projects — tragically, some of us found ourselves copying and pasting the card’s content rather than going through the process of moving it.
• Some cards are more useful than others — we need a way to indicate this because manual tagging is not enough.
• You don’t actually move from “jumble of cards” to organizing them on the page. There’s a step between having an idea and using that idea that we’re missing. Many of us found ourselves leaving the cards unmerged in the doc as long as possible, as our thoughts became more focused on the page.
• The name of the card owner doesn’t stay with the card when it’s moved onto the page. It should be captured and carried over like a citation is. In fact, we need more metadata to stay connected to a card.

Even in writing this post, I would have liked both the ability to mute used cards and to organize my thoughts / cards in a different sort of workspace before trying to write around them. By archiving the cards, I’ve now sent them off into my Archive Abyss, which is unfortunate because I’d like to keep all those cards in this project so we can address them in their original form, not my edited form.

In addition to robust search, muting cards and editing unmerged cards, these ideas floated quickly to the surface for us:

• We need a way to indicate that a card with repeatable / reusable content is reusable and that the others — the one-off snippet said by X person — is not. Perhaps a flag?
• People invited to a project will not always want that project’s cards added to their Card Library. We need a way to help people keep their Card Libraries clean.
• Card grouping!
• A small, medium and full view of the cards, in and out of the Card Library, would be great.
• …Spell check.
• Improved word count for the various parts of your doc.
• A headline generator could come in handy. Or a headline corrector / checker. Something to help bloggers come up with those oh-so-critical headlines.
• We need a way to wrap content in the document into a card, keeping content on the page as portable as possible.
• Make it easier to hide the Card Library and Image Library. The left-nav icons aren’t right.
• “Export to Medium.”

For the editor, tabs could get slightly unwieldy, as they do in Excel.

Outside of that, the editor itself performed very well. None of us wanted for additional formatting, although the usual stuff about image-editing came up — but nothing dramatic.

The Outliner disappointed some and saved the day for others

I personally rely very heavily on the outliner when I’m writing. I use it at the start of the writing process, to begin organizing my research and identifying gaps. I use it while I’m writing to quickly move a card from way down the page to way up the page. And I use it to move whole sections around. In the two hours I’ve worked on this post thus far, I’ve switched to Outliner view a dozen or more times.

But I’m evidently a power user.

Others found the Outliner to be “disappointing.” Said Sarah:

I’m using the Outliner to organize not only the general topics of my blog post but also thoughts / experience about the topic (while in the outline so as not to move b/w the outliner and document modes), but this experience is clunky and frustrating. I add bullets as a new thought (while in outline mode) but when I go to drop-in my research that corresponds to the particular topic heading it won’t allow me to place the research next to the bullet, where I want it.

Still others didn’t add headlines to the Outliner until the cards were organized, which is when themes would start to present themselves. This isn’t an expected use of the Outliner, but it’s a good one — and it shows the flexibility of the Outliner. We’re not forcing a way of working on people.

The Researcher really works!

Everyone used the Researcher (our Chrome clipper) when writing, and it became a key part of successful writing in Airstory. So that’s cool.

The challenges with our clipper are the same ones with all clippers, including the Evernote clipper:

• You can’t clip from PDFs
• Coding on some sites makes it tough to clip clean snippets
• Coding on some sites fully prevents clipping

Some challenges that we can definitely address include:

• Starting a project from within the Researcher — because sometimes you read something that triggers an idea for a post, and you don’t want to send your card to Unassigned, where you might never remember to go grab it
• The need for help using the Researcher from inside the Researcher (e.g., links to help videos)
• Quick-saves in the Researcher: hit ‘enter’ instead of clicking the ‘save’ button
• Bookmarking links instead of clipping sections of text from a page

All of those are quick fixes, or so I’m told.

And then there were all the other things…

Little stuff — like making project descriptions mandatory. Why do we do these things? When will we stop?

By this point I can’t help thinking: exactly how big do SaaS roadmaps actually get? We’re being very transparent here — I’d love to see what roadmaps look like for early stage vs growth stage.

New ideas to optimize kept cropping up for us during the month:

• Because ideas hit you anywhere and because we’ve got loads of commitments that keep us from prioritizing blogging (e.g., family, daily work tasks, meetings, exercise, errands), a native iOS app would be amazing. Steven and Sarah both wrote from their phones, and the responsive Airstory interface worked great for that. But the tab experience wasn’t great. And an acceptable number of clicks on your laptop is not an acceptable number of taps on your phone. You need to move really swiftly on your phone — and a native app could help with that. (As could optimizing Airstory for more shortcuts.)
• Dramatically better treatment of PDFs inside Airstory. By importing a PDF into Airstory, you can do much more with it than you can if you leave it online and try to clip from it there. But when you import a PDF into Airstory and clip from it, the card you create doesn’t have a citation (because that’s just not how in-Airstory clipping works). So there’s stuff we can do there.
• Offline mode! Three times this month, I was on a plane trying to write a post, knowing I’d need to rely on spotty wifi to get through it, knowing that would be impossible. I’m submitting this post late because my WestJet flight last night had no wifi. Offline will help members of fast-growing teams, where travel is part of the gig (likely for the reviewers if not for the writers).

What we didn’t explore much whilst dogfooding: collaboration

As the editor for The Better Story, I was the only person invited to a blog-post project. So the majority of the takeaways here come from me:

• We need browser notifications! When a project is ready for me to review, I need to know at that moment. Too often I’d go into a project long after it was ready OR while the writer was still working on it. I was busy guessing if it was time yet; Airstory should tell me it’s time.
• Both the editor and the writer need prompting when a deadline is looming — and they each need their own deadline. The writer’s deadline should absolutely come before the editor’s deadline; currently, Airstory has no solution for this.
• We need clean ways to stop collaborators from working in a project when it’s under review — and to keep reviewers out while the writers are still working. This will need to extend to the Researcher and other spaces: once a project status is changed, all the parts of Airstory need to respect that change.

We didn’t encounter the need for track changes — commenting did the trick just fine.

It would be ideal, however, to have strong statuses and notifications. Those are on the horizon for Airstory, but this need was really only felt by the editor (moi), which may be why it hasn’t come up that much among our one-person-team users.

How humbling to use your own product for a month. How awesome to know you can improve it…

Dogfooding isn’t about praising your successes, which is why this post is super-light on all the awesomeness of Airstory. It definitely highlights the biggest challenges and points of friction in your product. For us, we’ve got a big ol’ list of opportunities to add to the list users have already shared with us and we’ve otherwise developed. #CrossCountryRoadmap

The obvious bright side for us in discovering a lot of opportunities while eating your dog food is this: we can fix basically all of them.

Better than that? Eating our own dog food has given us stellar empathy for our users.

That, and we’ve got brand new insights into ways to go well beyond simply fixing the writing process. There are ideas we haven’t shared above that have got us buzzing with excitement.

tldr: Eat your own dog food. It tastes awful and it works.