Bottom-Up Architecture Q and A: Organizing the Site

This entry is part 3 of 7 in the series Bottom-Up Information Architecture Q and A

Once the reader reaches an Every Page is Page One page, does it still matter if the site is well organized? It depends on what you mean by “organized”.

This is another in the series of post dealing with questions from my TC Dojo Webinar on bottom-up information architecture.

Q: I understand that the site is not online, but it is the pages…however, you still need to have a site, right? …once a user gets to a page, don’t you want them to be in an organized site?

A: Absolutely, we always want our content to be well organized. But electronic media, and the Web in particular, have profoundly changed what it means to be organized.

In the physical world, organization means choosing one aspect of the relationship between objects as more important than all the others and placing those object next to each other and all other objects with the same characteristics. In a box or on a page, a object or a text can only be close to one or two other things. Organizing means choosing which things will be placed together, and positioning them accordingly.

In the digital world, there is no such constraint. You don’t have to choose one principle of organization over all the others. You can relate every object to every other object with which it has anything in common.


Map of a simple web

is just as much organization as this:

Content arranged hierachically, as in a TOC.

Consider sites like Wikipedia or Amazon. Neither has a table of contents.* Both have search mechanisms, and both provide extensive topical linking to related subjects. Generally, they do not feel disorganized.

One of the interesting things about purely bottom-up architectures is that good ones do not feel disorganized when you are using them for a specific purpose, but if you step back and ask yourself “how is this material organized” you discover that there is no external “organization” that you can look at outside of the content itself.

This can be disconcerting, particularly for a writer. But consider that with a top-down architecture, while it is always possible to step back and see how the whole is organized, you can still very easily feel lost when trying to solve a particular problem. Knowing the structure of the content collection is not the same thing as successfully finding the subject matter you care about.

The only time that a reader is likely to step back and ask the question about how the whole is organized is when they do feel lost in the content. It is possible they may find relief in a TOC at that point (which is why you should not begin a move to bottom-up architecture by throwing out your TOC). It is also increasingly likely that they will go back to search rather than ever looking at the TOC. But with a good bottom-up architecture that links along all the major lines of relationship in the content and its subject matter, readers are far less likely to get lost in the first place.

Organizing the content has always been a particularly sensitive issue in the book world. In a book, you have to choose exactly one point in the linear sequence to insert a particular topic. Every page is before exactly one page and after exactly one page. Choosing which two pages those will be is a complicated business, and the result is always a difficult compromise. This is what makes organizing content a particularly important skill in the paper world, since it involves these hard choices.

But online, you can place any page next to the current page simply by linking to it. A page can thus be next to a large number of different pages, without any necessary subordination. Every page that is linked to is an equally possible “next” for the reader.

Organizing content in this environment involves far fewer hard choices. You don’t have to pick one next out of all the possible nexts. You only have to provide links to all those possible nexts. That is what organization means in a hypertext environment.

This is not to say that organization is less important in the hypertext world, only that it does not involve the same hard choices between rival forms of relationship that are necessary in the paper world. In one sense, it is a more intense business to organize content for hypertext, since all of the important relationships are still in play. With paper, once the hard choices were made, the losing relationships could often be ignored. With hypertext, they all should be captured and expressed.

Of course, hypertext organization falls apart if those multiple next pages don’t actually fulfil the promise of the link that leads to them. The linked pages need to be correctly identified and they need to be effective Every Page is Page One topics so that they actually do the job the reader needs them to do. As we would expect in a bottom-up architecture, organization of the content set starts at the bottom, with the individual pages.

* Actually, the sidebar of every Wikipedia page does have a link to “contents”, but what it leads to is not a table of contents, but a list of lists. Lists are an important part of bottom-up information architecture, as are lists of lists. Wikipedia even as a list of lists of lists. Lists may be stand-alone or they may be embedded in topic pages (every page is a hub of its local subject space). Lists are organized by subject, not by the shape or boundaries of the content set. Lists can respect and reflect as many types of subject relationships as you care to create a list for.

Series Navigation<< Bottom-Up Information Architecture Behind the FirewallThe role of the TOC in a bottom-up information architecture >>

, , , ,

2 Responses to Bottom-Up Architecture Q and A: Organizing the Site

  1. Rebecca Hopkins 2015/02/18 at 12:18 #

    I think Tech Pubs will never get away from the TOC or some kind of structure because we have to organize our site somehow. We’re not computers able to grok a mindmap of a thousand pages. The second graphic must be considered more human-readable, more user-friendly, and with inline links, the second graphic becomes the first one.
    A TOC is a place to start, and maybe the last resort if search fails, but the vast number of users may not care about it. Tech Writers do a lot of stuff in the background that is transparent to the user (or should be) so let writers have their hierarchy and the users can search and cross-link to their heart’s content.

    • Mark Baker 2015/02/18 at 18:35 #

      Thanks for the comment, Rebecca.

      Indeed, writers are always going to need some sort of structure to do their jobs. Readers can proceed by information foraging rather than by following a structured plan, but writers can’t work that way.

      Furthermore, while the bottom-up structure of a bottom-up architecture is just as much a structure, and while it works better for an information foraging reader than a top-down structure does, it does not do the same for the writer.

      Writer’s need some form of structure that they can stand back and look at as a structure, but you can’t stand back and see the overall structure of a bottom-up architecture because the structure is created by the topics themselves and how they link to each other.

      Writers who understand and appreciate the power of a bottom up architecture to serve reader’s needs can still be frustrated by their own need for structure in their tools and processes.

      Creating a bottom-up architecture is not difficult. A wiki is a great bottom-up authoring tool. But wiki’s don’t have a lot of inherent structured management features. Wikipedia gets around this through the constant work of thousands of contributors and editors. No one person sees and governs the whole, but the whole is well managed through the cooperation of the crowd.

      Crowd management is a harder thing for a small group of writers in a pubs department to pull off. It can be done, but it takes a new discipline and a new way of organizing and managing your work.

      The alternative is a structured approach. If your content is structured (in a way that supports it) you can use algorithms and metadata to create and maintain the structure of your content set. If you need to visualize that structure, you can use queries to see different views of how you pages are related to each other. This approach is unfamiliar to most authors, but it is common in the database world.

      With a new information designs comes the need for new processes, new tools, and a new way of thinking about and creating structure. We need to think more about what needs to be done in order to create a structured environment for authors to create bottom-up information architectures.