Why website Information Architecture is the ultimate Digital team sport

A thing cropped up at work, at the time it felt a small thing, but as I posed a question in Twitter it soon became clear to me that (a) I had very strong views on the matter and (b) more interestingly, my views were quite divergent from colleagues and friends. I thought I’d write things down, either to help myself, or to help my colleagues better understand my viewpoint.

The question I asked at work (in a room of 20 people) was “who owns IA?”. As I’ll say at the end, this was the wrong question, but it stimulated a lot of views and caused a lot of reflection on my part.

What do I mean by Information Architecture?

When I say Information Architecture (IA) I mean the art and science of ‘town planning’ for a mega-website.

I admit I can be a bit old school. I built my first website in 1994, I went on to embrace social media management in the 2000’s, was custodian of public sector websites that served 11,000 staff, 200,000 students and took £200M per annum of ‘sales’ per year. That was at the UK’s biggest University where faculties were federal states, and the views of the Marketing Director were not necessarily aligned with the Pro Vice Chancellor for Research.

This meant the purpose of the website(s) wasn’t agreed, which meant creating and managing an IA was a challenge at best. That challenge brought with it really awkward things — like who and what should be on the home page, who should have what URLs (including vanity URLs) and how would we measure and report on success.

I now work in Health — and have worked in an Agile fashion and as a Product Lead since 2012. Part of my work involves the transformation of a mega-website — mega in at least 3 dimensions: c. 50M sessions per month (30M+ unique visitors), anywhere between 5,000 and 25,000 webpages (depending on how you count) and most importantly content and access to services that saves lives.

I know we should be talking about Service Design — and I can — elsewhere I talk about a Service Standard for Health that encourages teams to “join things up and work towards solving a whole problem”, but right now we just need to sort out the IA of this mega-website, and make sure our content is structured and tagged so that (a) it can be surfaced (syndicated) everywhere and (b) we can then design, build and integrate service based journeys for the future.

The twitter meme

In Twitter I thought I’d see what the crowd thought by asking “is the skill of doing Information Architecture a design task/skill?” and kind of gave my view away by adding “My view is not…”

“Asking for a friend — is the skill of doing Information Architecture a design task/skill? My view is not”

My design colleague @deanvipond replied

“I see IA as bringing understanding and order to messy, complex problems. So yes.”

And my (other) design colleague @karlgoldstraw contributed:

“IA is a tool of the design process (aka ‘UX’). No one person can be responsible for it and it requires a plethora of human input to get right. But you’re still taking a thing, chunking it, and making it better, and that’s the very definition of design process.”

Then @cjforms introduced a sense of scale:

“Need for IA in design might go like this: Designing a chair: not so much. Designing a chair that will be part of a range of furniture: possibly. Designing a chair that will be sold through a catalogue: probably. Designing the catalogue: definitely”

Which @NicolaatDH illustrated beautifully in highlighting IA as a classification issue:

Twitter image showing a furniture catalogue.

And @jesmond brought together the idea of design, and spreadsheets!

“I would suggest that no-one could design said IA without the aid of a spreadsheet. (As well as a consignment of post-its and a large wall) But they’d still be designing an IA”

And my product colleague @jiggott brought wisdom to bear by naming specific skillsets and the need for an IA specialist:

“Are you asking which profession/tribe information architects belong to? I would expect most designers to be able to solve simple IA challenges, but when you get into large, complex taxonomies and ontologies you want a specialist IA.”

And @alastac and @emmachi shared their stories of apocryphal spreadsheets (which I consider very useful if doing ‘proper’ IA work…)

“We once did an IA project where we hit the limit of (32 bit) excel, which was about 32,000 rows.”

“As head of IA at Pearson I wrangled a 5000 row spreadsheet of taxonomy & ontology and a hefty dose of research, the navigation design ended up with UX and it caused a few butted heads but the disciplines were kept separate (even though I’m a dual specialist)”

(The whole thread is fun for a wide range of views. Have a read if you have chance)

Highlights from the Twitter thread show the wide range of skills and views on what IA is. Yes, Design is involved (of course) but so is User Research, Data Science/Analysis, Library/Information skills and Tech. Quite involved then…

A definition of ‘doing’ IA

So I thought I’d try and define what I mean by ‘doing IA’. I believe ‘doing IA’ is evidenced by:

• Having aspects of data science or library information skills available. This includes ‘categorisation’ which means understanding taxonomy (the classification of things) and ontology (the properties of a thing, and the relationship between things)
• Understanding the importance of standards — people use metaphors of IA being ‘town planning’, or like architecting a house. If so, then you’ll understand the importance of a standard gauge for trains (an issue in the 1800s) or that your plumbing pipes are all 12mm or 15mm and you have a ready supply of adaptors to make things fit. In the web world the equivalent is common components and patterns, navigation conventions and things under the hood re how data is shared/login preferences preserved.
• Understanding IA is there to meet user needs (which equate to business needs) so you need to be able to measure that you’ve met those needs — not just in user testing but through effective web reporting with the ‘pitter patter’ of millions of feet across your web terrain. Web measurement tools like good logical structures to measure against, which makes it easy for digital teams to set up reporting which is easy to read and understand.
• Having a logical IA tree, and I don’t mean 2 deep. I mean as deep as you need right now, but also knowing how it will grow for future scenarios (which is where a spreadsheet is handy). This also means a ruleset so when you have a new demand or a really awkward request you know where it sits (think dewey decimal). Otherwise you have a section for FAQs or the equivalent of ‘miscellaneous’ as one of your main headings (which is very bad).
• Having a URL naming policy. Deciding on your URL structure isn’t a technical thing, it’s a business one that aids your content strategy and your visibility in search (SEO). It’s also a key governance issue and point of decision making. It also allows you to say ‘canonical URL’ frequently.
• Semantically structuring everything. Both the format of the page (e.g. H1s, H2s), but also ideally using a subject based schema (schema.org) so you can ‘chunk’ your content and open up the world of syndication, open or linked data.
• Understanding the key role of search. Partly because your site search supports when your IA fails, but also because most of our traffic comes from Google. Structure, logical URLs, non-conflicting content, appropriate tagging, and volumes of visitors all support excellent search visibility.
• You need a search policy — to describe what good looks like for your organisation (i.e. a target for how you want to appear in Google) but also to describe what you want to include in your site-search collection, and how you want your own competing factions to appear.
• Having the right tools and techniques available to design and test an IA e.g. site mapping tools, treejacking type software, card sorting exercises and ways of analysing data (which is the other place where the spreadsheet comes in…)

Modern ways of doing things…

Over the last few months when I think of IA I’ve wondered how it fits with the ‘modern’ post-it led way of doing things.

A wall of post-its

Now I love a good post-it session, we’ve filled many walls with them, but if what I have said above is even half way true I don’t think you can avoid using additional techniques and doing stuff that might seem old school to deliver an IA strategy for a mega-website. IA is difficult, it is gnarly, and it requires particular inputs. I don’t believe post-its alone, or prototypes (of whatever fidelity), will get you to a deeply thought through IA strategy for a mega website.

Mapping the IA task to digital teams

If you have a multidisciplinary team in the way that many public sector bodies think (see UK Gov’s Digital, Data and Technology Profession Capability Framework) then I think each role brings something unique:

• Product Owners need to be a jack of all trades for IA (i.e. a bit of knowledge of all of the above) for two reasons — to shape the ‘definition of done’ and to ask the right questions of her team (and to reflect that many Product Owners see themselves as servants to the team).
• User Researchers think about the techniques to discover and test an appropriate IA — and once you have some insights is able to make sense of the data by analysis — which may include using spreadsheets! Even better if you have a data scientist…
• Designers are natural problem solvers, designing solutions (patterns, nav structures) and may stretch into a visual sitemap, if she’s got a big enough screen….
• Content Designers understand how structure (‘on page’ and ‘across site’) is vital to content, and how journeys help grow comprehension.
• Developers help ensure code supports a meaningful IA (semantically true, marked up)
• Architects make sure the right tech is in place to form user friendly URLs, to use ontology (servers), to deliver structured (and efficient) mark-up (including via API), and provide technical sitemaps to help search engines etc.
• Delivery Managers make sure all elements of a successful IA are in place, and marshal across content/design/UR and tech to make it so!

Back to that IA question….

The question I should have asked at work is ‘who is the Guardian* of the IA?’, because working in the public sector (as @cjforms pointed out) the owners of our services are the tax paying public, and in my case people who are patients, supporters, carers and people who work in health.

So it’s a bit of a cop out to say ‘everyone’ in a digital team is a Guardian of the IA of a mega-website — but it is why I say Information Architecture is the ultimate digital team sport, and cannot be laid at the door of just one of the professions.

Ps — you can also read ‘Why website Information Architecture is a metaphor for Digital Transformation at scale…’ which takes IA thoughts and puts them into more of a strategic context

* Guardian — like the Caldicott Guardian is a much better phrase, implies a weight of responsibility, privilege and the ability to be able to provide leadership and informed guidance on complex matters involving confidentiality and information sharing. (though I will still remind there is always a responsible owner in terms of who would end up in court if things go wrong, so the Guardian is also sometimes Police Person…)