| I read a post recently that illustrated how concept models are rarely used in the right way and are often misunderstood. Are they really worth doing at all?
Now seems a good time to expand on the tool that Dan Brown has popularised through his book Communicating Design. Not as simply a stand alone tool but one that can provide a blueprint for giving solid ROI on design, analytics and testing.
This sketch shows the tools we use at Hello in our UX team. Not every project warrants it but the one item that binds so many tools together is the Concept Model (CM), sitting between Analysis and the Information Structure phases. It describes the big picture, the unique proposition of the site that separates it from its competitors.
Add the users into this and you have a solid core of basics mirrored in the principles of IA. Essentially it places the user at its centre (user needs), the business interfacing with the user and its own objectives (the context) and the actual content that will be produced often in the form of a product (content).
You should also blend in real world statistics to give credence to your thinking. Many business owners like to be shown a few figures and it will back up your argument and it also makes for a more in-depth read.
By giving a layered concept model it becomes a document born through the emergence of factors evidenced through research.
A site map’s foundation
By using the model as a basis for site structure, the site map can be spun out fairly rapidly and its presence ensures the user needs, the business goals and content types are never missed in the design process. (A quick comment about site maps:- some people no longer use them but the client requires it in my experience, and helps them gain understanding.)
I find that the CM is the most important of the deliverables set, as it crystallizes the thought processes of a team and defines how a project will be tackled and what the likely outcomes will be. All with the user at the centre of the design process, based from good qualitative and quantitative research data.
Does the content fit the concept?
Once the initial concept has been built by aligning research findings with the business goals and the context of what the product will provide the user with, we can move on to more detailed aspects of the model.
As we define strategy quite clearly from the model the next thing to do is define where there are gaps in the proposition. Is what you are intending to supply or offer as a service meeting the requirements of the customer/user? Does the experience that they wish to have from using you, satisfied by your offering?
Drill down to each area or node and see what is missing, by using multiple layers or pages you can place different elements on the same map to give a visual overlay that allows the viewer to see the gaps or the areas of strength. It becomes a visual gap analysis/SWOT snapshot.
Some things may be intangible or not even content types but are emotions you wish the user to have, or interactions that they need to engage in.
Prove it works
This iterative way of revisiting the original concept from a content angle can enforce changes to the initial vision. But this is a good thing as it is making the model work harder and enables it to prove that it is a winning concept.
It’s a good idea to show and communicate this diagram with as many stakeholders as possible. Their views will be a test for the model; it will either change or stand up by itself. Testing rigorously at this stage ensures that when the physical build occurs, certain assumptions or hypotheses have already been checked.
Engagement overlay and emergent structure
The next stage is to employ methods that could be termed as action research. It is seen as;
‘…beyond reflective knowledge, created by outside experts sampling variables to an active moment-to-moment theorizing, data collecting, and inquiring occurring in the midst of emergent structure.’ (Wikipedia)
We can do this most effectively using web traffic stats and a skilled web analyst. With informed analysis (using personas) and clear measurements of user engagement we can build a true picture of visitor behaviour. Omniture and Future Now (Bryan Eisenberg) have a great white paper on this technique here
On an e-commerce site an analysis of the funnel is clear. However on sites where content is consumed but no transactions occur we must use a different set of metrics to gather if our visitors are consuming pages or interacting with our content.
The context of these statistics must always be carefully gauged as it needs to appear as a balanced view of the user population. By determining bounce rates, page views per person, returning visitors and those who subscribe to services or purchase items we can build a picture of different user behaviours.
We will always see a different behaviour from SERPs than from RSS visitors for instance. There are different qualities too if they have stumbled upon an article to those who have a need to be informed. It is important to conduct the test in a live environment that the information gathered has the ability to change the concept. The model needs to be fluid and adaptable, and not be bounded by prejudged stereotypes.
Personas can help here to give us an idea of what the background of the users are but an extra facet is that the metrics inform of how those actual users behave. This is an ‘emergent structure’ mentioned above, the unknowns of the audience become clear from this phase of analysis. We do not know what the users will do next, only what they have done and what is emerging.
Testing the model in an iterative process
Once the diagram has been interrogated by stakeholders and live statistics are also placed alongside the content propositions we know what we have is a concrete strategy. From here we roll out the production of the wireframes, based on the sitemap and taxonomies and other deliverables.
Here the model can clarify interactions. By looking at the purpose of the functional elements of the design alongside the strategic goals of what the product will offer the user, the CM defines how a user wishes to interact with the product.
As soon as we have our prototype we can then test the designs in the live environment using A/B or multivariate tests. Measuring how the users engage with the site at this stage allows for optimal designs that can be designed and refined iteratively as each test result is analysed.
Modelling real agility
This way of developing pages fits well within the agile method of development. By using an iterative testing model aligned with the sprint cycle of development, changes can be made to designs quickly. Testing incrementally and doing it often is a great way to measure how engaging the pages are.
At this stage, look at the prototype and see if anything has drifted away from the concept model in terms of what the site delivers. You should be able to look at the site and the model and see parity there. Key elements of the concept should have been translated on screen and the interaction, stated on the model, flows from wire frame to the html prototype.
Maintaining the vision
This way of sense checking is an easy task with a model that is circulated and pinned up on the walls of the project team. At all levels this deliverable is relevant. The nearer the project gets to release the more important it is that the CM is referred to. It is a keystone and a point of reference. Too many times the build of a site overtakes the overall vision, here it should help to ground the team and provide a common focus to the overall goal.
Concept models as a conversational tool
CMs detail the big idea that needs to be followed, beyond intricately detailed site maps or the interactions of a wireframe. They are beyond taxonomy and personas and yet they encompass all these things in their make up.
At their core is the reason, the business case, for their existence. It is this fact that enables all stakeholders, CEO’s to web developers, to understand why they are so important and communicate the ideas that they portray.
The real strength of these diagrams is their ability to cope with the emergence of user characteristics during analysis, and that they have the flexibility to show this.
So before you design or re-design perhaps the best thing to do is draw up a concept model, grab as much existing research as possible and get friendly with a web analyst. It will be the most powerful deliverable you will have in your toolbox and provides real ROI for your client.
Concept Models - the underpinnings of design ROI…
- State the vision – They allow teams to have a shared vision of a project that is easy to understand and communicate with each other conversationally
- Defines structure – They help dictate the structure and organisation of a site by detailing the components of a design
- Defines content – They allow content to be tied into explicit user requirements and business goals that are reflected in the model
- Defines interactivity – They help define interactions that a design will need to display to satisfy the user and enable it to fulfil its objectives
- Enables rigorous interrogation – Measuring the engagement of the user base in the live environment enables a rigorous checking of the model, and the prospect of the site succeeding can be tested before real development work begins. From the results, it effectively gives context to the project
- Plants an anchor – The model can also be used to ensure that the development work does not drift away from the original aims of the project, that the completed design fulfils the business goals and the user needs, serving up the content in the best way possible.
- Offers project flexibility – The concept model is adaptable and can be readily changed and has minimal cost implications if a change becomes necessary.
Tags: Agile, Collaboration, Concept Models, Design Practice, Diagrams, Experience Design, Information Architecture, Information Design, Mental Models, Mind Maps, Taxonomy, user, User Centred Design, Web Metrics, Wireframes