It’s Information Architecture, not Content Architecture

A couple of months ago I kicked off World IA Day at General Assembly Melbourne with a talk called “It’s Information Architecture, not Content Architecture”.

I started by talking about how dealing with content and IA at Envato allowed me to spend a lot of time analysing the category structure of our marketplaces. I wanted to understand the system, and identify its key pain points, in order to improve it.

I gave the audience some context about Envato that would be particularly relevant:

  • Envato mainly builds and sells digital assets and services for creatives,
  • It’s a bootstrapped company that literally started from a garage in 2006,
  • It’s now a global company that has grown exponentially in just a few years,
  • Our marketplace for website themes, ThemeForest, ranked last year among the top 100 most visited sites in the world (according to Alexa).

The first marketplace founded in 2006 was FlashDen (then renamed to ActiveDen), selling Flash templates. After its success, Envato saw the opportunity to tap into other markets and built a total of 8 marketplaces over the years. This eco-system of marketplaces, Envato Market, sells digital assets created by a worldwide community.

We recently bid farewell to ActiveDen (because Flash, lol), but the other 7 marketplaces are up and running. You can find basically everything you need for any kind of creative project. From website themes to stock video, from royalty-free music to vector images, and so on.

The success of our marketplaces has been massive over the years. Goes without saying—this meant that the amount of creative content that our authors uploaded to our sites has grown exponentially. Today, we sell millions of items to millions of users all over the world.

To put it into perspective: we went from 50k items at the end of 2010

…to a whopping ~10,000,000 in April 2016.

Yep, ten millions. I mean, that’s a lot of stuff.

And of course, when you get heaps of new content coming through, you want to put it out there for users to find it. And the easiest way to do it? Well, for a few years it seemed to be to just create new categories. As new trends, technologies and types of content started rising, bam! Just add a category and that will do.

So I asked my audience: Does it sound right to you?

Well, it shouldn’t. Because this is kind of what we ended up with:

And I mean a freaking scary mess.

Categories on top of categories, duplicates, categories that could easily be replaced by tags (i.e. Nature, City, Buildings) or other metadata, unclear naming and labelling that would make sense to us but not to our users…

This led to sharing with my audience what I think is a valuable lesson I learned:

Information Architecture is ever-evolving in the world of e-commerce. Trends and markets change, new products appear. Your IA needs to reflect that dynamism.

If you can’t keep up, your users won’t discover the content they are looking for on your site—even if you do have it. They just won’t find it, and will look somewhere else.

I then provided an example to introduce a couple of key learnings. I showed the audience a piece of stock footage that you can buy on VideoHive, our marketplace for all things video, from templates to footage.

The video is basically just an animated version of what you see in the screenshot above. A fishing boat in water.

It belongs to the main category “Stock Footage”. I then listed 3 of its many subcategories: Nature, Water and Vehicles. I then asked my audience: Which of these subcategories do you think the video falls under?

Some answered “Water”, most answered “Vehicles”. No one chose “Nature”.

But as the following slide showed…

Yep, the video currently falls under “Nature”. And yep, they laughed.

I left the issue of authors picking the wrong category during the upload process aside, and focused on another key issue.

Categories should be mutually exclusive, i.e. you choose one over the other.

So with that in mind, I asked myself—why did we create those categories in the first place? Imagine you have a video of a motorbike riding in the woods, next to a lake. All those 3 pieces of information are relevant. Why should we pick only one?

(At this point I also reminded the audience that, due to a technical constraint in the back-end, items cannot fall under more than one category in our marketplaces. So this mutually exclusive thing is particularly relevant to us).

Perhaps these 3 instances shouldn’t be categories at all then. They look more like tags: labels you can apply to the same item to describe it better.

I then went on to share another very important lesson I learned from my work with IA.

Information Architecture is not just about categories.

It’s very easy to identify Information Architecture with categories (especially when you’re in a hurry!). We forget that we have other ways of capturing information about our content and present it back to the user in a meaningful way. I’m thinking about tags, attributes, and meta-data in general.

Information Architecture is about how you capture information and how you go about making a sense of it.

If there’s confusion about what should be a category, and a tag, and an attribute, and how they interact—you’ll get to the mess I was talking about just before.

So what does this mean to a user?

Creating a good Information Architecture ultimately means defining the discovery patterns that users will follow to find content on your website. Patterns that need to be predictable, repeatable, easy to understand.

I then went on explaining how a good Information Architecture helps users understand:

  • Where they are,
  • What they’ve found,
  • What’s around,
  • What to expect,

And how it’s important to plan ahead and define “what is what”:

  • What information do we want to capture?
  • How are we going to capture it?
  • What are we going to treat as category?
  • What should be a tag instead?
  • What will be treated as an attribute?

And then reminded them that there’s one key thing to always keep in mind:

The reason why I decided to give my talk this kind of provocative name is that my experience has taught me this very important lesson: You need to start with the information.

If you are able to nail the way you organise the information, i.e. what is a tag, what is a category, and what is an attribute, the rest will fall into place fairly easily. And in the meantime, you’ll have created in the back-end an architecture that’s flexible enough to enable you to present the content to the users any way you like (read: the way that’s more valuable to them), without constraints.

I then ended my talk bringing back the audience’s attention to the users once again. Ultimately, all this defines the success or the failure of the discovery patterns you’re creating for them. And if the pattern is “OK let’s see what the hell happens if I click here”, well… UX issues will be coming your way. And visitors will be doing the opposite—heading to your competitors’ sites instead.

I was really happy to be given the chance to do this talk, and got very good feedback from the crowd. I also enjoyed the following talk by my fellow Envatian Kate Hunter, SEO guru, talking about how IA impacts site ranking on search engines.

And that’s me with my Macbook full of stickers.

Show your support

Clapping shows how much you appreciated Fiorella Rizzà’s story.