Why sway.com (still) sucks

sway.com is a new product by Microsoft that should make publishing easy for everybody. Of course we're interested when one of the big players gets his hands on an online publishing editor, after all that's what we do too with livingdocs.io.

First things first: the sway editor is way too complicated. Entering the title is still frictionless but that's where frictionless stops. Sway takes a completely different strategy compared to other online editors, say medium.com, in that it breaks the page into little cards that form a storyline. This is not WYSIWIG. Instead the strategy strongly supports structured content. But lets take the sway editor apart piece-by-piece.


Text is entered by typing in one of the cards. If you press enter then a new card will be created, similar to a new paragraph in a classical document. The formatting options are always shown at the top of the page no matter where you are writing text. This is a bit confusing and less intuitive than the selection popover in the medium editor. Furthermore, the actions triggered by these options, e.g., the "header" option, are sometimes completely intransparent. Selecting the "header" option will not only make the selection big (which could be assumed) but also move it to another card (which is rarely intended). What makes these kinds of interventions even worse is that you cannot undo them (Ctrl+Z does something weird but it's not undo).

New cards can also be added by pressing a little plus-icon at the bottom of a card. This seems intuitive. What is very confusing is that the plus-icon only appears on the currently selected card or the last card that was selected (if none is currently selected). Why should content only be added after the last selection? A better and often-used solution is to show the plus-icon on mouseover between two cards (or paragraphs).

A title card and a text card in the sway editor. The "add more content" plus-icon is on the last selected card.


Sway offers a whole lot of elements that can be used in a story. After clicking the plus-icon you can add a generic "Media" element. This gives you a list of different sources such as Bing, Twitter, Youtube, etc. In each of these sources you can search for your content via a search input. While this works perfectly for the source Bing (a search engine), it doesn't work at all with Twitter. The reason is simple: you don't search tweets, you follow them. So if you have a great tweet in your stream there is no way to get it into sway. You can not drag&drop from Twitter itself, you can not go via an embed code and search for it by keywords is close to impossible since the search is so bad (I tried searching for my Twitter handle getting no results). The weirdest thing is that they did it right in the case of Facebook (you connect with your account to get your stream), but not for Twitter. Maybe Microsoft doesn't like Twitter?

Different kinds of element sources for sway. The example shows a Twitter search for "javascript".


Publishing to different channels.

Sway offers the usual publishing channels such as "link to the article", "tweet it" or "send it to your Facebook friends". Well, at least that's what the buttons indicate. Clicking the link and embed buttons actually does nothing (no, there is nothing in my clipboard). Also, I always see a message telling me "Need to revoke access?". Do you want to get rid of me Microsoft? Things like these make sway feel pretty buggy.

You can preview your page by either selecting a preview button or just pressing to the right of the cards. The result is a little weird. Sway automatically arranges the content of your cards in a layout. You can change from a pre-defined set of layouts but you can not decide per element what goes where. The same is true for styles and colors: you just select some presets and sway does the rest for you automatically and to be honest not very intelligently. The remix button tops this: it just selects some combination of layout and styles randomly. This is a toy, not a feature.

Bottom-line for the publishing workflow: you throw your content at sway and it will do something with it. You have very little control over what that something is. While this might be ok for putting online some holiday photos sidelined by smileys, it makes sway very inadequate for people doing serious work. I might be wrong on that last point though since I heard rumours that there are even people who use PowerPoint for serious work.


Sway brags in its video a lot with its support for mobile devices. Unfortunately, I am not impressed. I tested on a Nexus 10 device so maybe the support in Microsoft's mobile devices would be better.

The first thing that strikes very odd is the drag&drop. If you search for an image in the elements sidebar then you cannot just drag it to the page. You first have to select it with a tap, then it gets highlighted green and only then you can drag it. That's very non-fricitionless.

Synchronization between devices is also not very good. When I wrote simultaneously on my laptop and the Nexus device then I got a message that the page has changed and I should refresh my browser. So, apparently they do have a push-connection to my browser but instead of updating the page intelligently like e.g., Facebook does with the stream, they require a refresh. Feels somehow unfinished.

The good parts

Are there good parts to sway. Certainly! The concept of structured content is very interesting. While the remix button is outright stupid it shows what could be possible: Structuring content intelligently with respect to output channels. Sway knows what a paragraph or an image in a page is and has thus much more control over how it can combine it. This is for example interesting in fields such as structured journalism that re-use content parts to increase the cost/revenue ratio of content.

The marketing is also very good. The video is inspiring and the content examples look all very nice also if slightly too fancy. I wouldn't know though how to produce these results with the editor I've seen...

Also, it's a good sign that Microsoft didn't build Word for the browser but realized that web publishing is a different beast. With a little (or also a lot) of work sway could become quite cool. Lets hope they are ready to make some bold decisions.