Here are a few tips based on our experience of user testing mobile ecommerce sites:

Meet expectations

Don’t disappoint visitors by removing key functionality or services they have experienced on your main site, e.g. reducing the delivery or reservation options. You may have to rethink the design to achieve this, but rethink rather than reduce.

Context of use

Many users have told us that they use their smart phones to check prices when out shopping or to look at reviews of a product. By understanding customers’ behaviours you can support them, e.g. barcode scanners for quick product price comparisons. Ask yourself what you can provide visitors in the different situations mobile opens doors to.

Data entry

“I’d give up now and go on the computer” is what we hear from users faced with too many input fields. Look for ways to reduce the hassle of data entry whenever possible, e.g. the facility to store details to make it easier next time, and providing options with check boxes wherever possible.

Data allowance

“Going back through so many pages is OK on WiFi, but what if I was using my data allowance?” Don’t frustrate customers with unnecessary page loads. Think about interactions that will be most economical.

Security

We find users tend to worry more about security when transacting on mobile sites. Provide appropriate reassurance and link to more information.

One account

A customer’s trolley and address details should be saved no matter which device they login on. Customers might add products to the trolley on a mobile but complete the transaction on a PC.

User test

Do it early and iterate. Paper prototypes, wireframes, interactive prototypes and live sites can all be tested. You’ll learn lots about your customers’ behaviour, how to improve their experience and increase conversion.

Want to know more?

If you’d like to know more about testing on mobile and tablets then why not get in touch with us – drop us an email (more@bunnyfoot.com), a tweet (@bunnyfootsays), whatever suits you.

Card sorting. It’s really as simple as it sounds, but one of the most effective tools in user centred design to understand how your customers navigate to structure your content and name your sections.

I’m going to use this blog post to try and give you an overview of card sorting – what it is, how we use it and why you should make it a part of your next design project.

Why does it matter?

Let’s take a look at an example:

Thom is a handyman and it’s his first time he’s been asked to install a burglar alarm. A friend has told him to choose “Safe as Houses” alarms, so he went to their site. He sees a list of alarms, arranged by code number with a heap of technical information. He just wants to know if it’s suitable for a domestic installation and if it fits his budget!

You see, the problem with building a site structure from within the company is that you know your own products too well. The structure you present to your Managing Director is very different to the one the customer would produce. To provide the best customer experience – navigation, structure and usability must be based on customer needs.

What actually is card sorting then?

Right, yes, back to the main point: card sorting is a user testing technique that splits your top content out onto cards, ready for your customers to sort it into categories that are intuitive to them.

The basic process of card sorting is as follows:

1. you do an audit of the key content that will be on your site
2. you create a card for each section
3. you present the stack of cards to a participant and ask them to sort them into groups that are meaningful to them.

You can do this:

• with a person present to capture qualitative data (moderated),
• online, or without a person present (unmoderated),
• with set headings (closed)
• allow the flexibility to make their own groups and headings (open)

Moderated, open and unmoderated online card sorting styles

How many cards should you use?

Given that the point of the card sort is to see how users would expect your content to be organised, each card represents a piece of this content.

We aim to have no more than 70 cards to sort if the items are easily understood, familiar items or 40-50 if the items for sorting are more technical and less familiar for participants.

If you have a lot of pages or site content that are spread across several levels of pages you might find it more suitable to sort 2nd or 3rd level landing pages rather than the deepest level child pages.

What to put on your cards

Typically, a card sort card will have a label that represents an individual piece of content (e.g. the name of a page) and often a short, non-technical description of the item on the back.

Depending on the nature of the items that need sorting and the audience doing the sorting, you might like to use pictures instead of or alongside the labels. Using pictures also has obvious benefits when testing with participants who have various levels of proficiency in a primary language. This might include participants of different nationalities, participants with learning difficulties or even child participants.

What does it all mean?

The power of card sort analysis is finding the correlations between participants – which cards are consistently grouped together, which are sometimes grouped together and which are never grouped together across all of your participants. These correlations can then be used to derive a proposed content structure for your site based around the way that your client’s users see the content belonging together.

What type of card sort is best?

Choosing the right type of sort for you will largely depend on the requirements and constraints of your project. If, following previous user testing sessions, you also have a fixed high level site structure and you wish to know how to organise content within it then a closed sort would likely be the most appropriate for you. However, if every option is open to you and you have no constraints as to the organisation of the content on your site then an open, moderated sort will offer you a wonderful opportunity to organise that content around your users’ expectations.

A good strategy for card sorting is to plan for two sorts at different stages in your project development plan.

The first would be an open, moderated sort and would take place once you had worked out what content you will have on your site. Conducting open, moderated sorts with 10 or so participants provides you with an opportunity to not only find out how to organise content, but also to find out if that’s even the content users would want to see on your site.

Once you feel confident that you have the right content and a user-centric means of organising it a second, closed and unmoderated sort should be conducted to verify the appropriateness of your site structure and content organisation. This sort would present 20 or 30 participants with each of the groups that you came up with following the first sort and ask them to place your content cards within those groups.

It’s often the case that a vague label or technical description has led to a card being difficult to group, resulting in inconsistent categorisation (evidenced by a weak correlation with any particular group). A round of closed, moderated sorting gives a further opportunity to gain insight into users’ understanding about those cards and reveal what needs changing about them in order for them to be easier to categorise.

So there you have it – card sorting in a nutshell.

Hopefully having read this post you’ll feel that you have a better understanding of what card sorting is (the sorting by participants of cards that represent content on your site into groups), why it’s a vital technique (you give your users the chance to guide the organisation of content on your site into a structure that is meaningful to them) and how and when you should go about it (it isn’t just for website testing, you can do it for in store layouts, apps, tablets and more).

Want to know more?

As ever, if you have any question about card sorting or would like to find out more specific details about how you could include within your next project then please do get in touch and ask away. Either drop Eleanor an email (eleanor@bunnyfoot.com) or grab her attention on the tweet-o-sphere @bunnyfootsays– we always love to hear from people so make our day and say hello!

Further reading

Card sorting: a definitive guide
Donna Spencer and Todd Warfel, April 2007

Card sorting: How Many Users to Test
Jakob Nielsen, July 2004

Card sorting: designing usable categories
Donna Spencer, April 2009
Rosenfeld Media. ISBN: 1-933820-02-0

They are a key part of ensuring you stick to your user centred design principles. This post will hopefully give you a bit of an overview of what a persona is and isn’t, how we’ve used them and why you should too. There are also some suggestions for further reading at the end if you’d like to get your teeth sunk a little deeper into this topic.

What are personas?

To the main point then and probably the question you’d most like answering: what are personas?

In essence, personas are simply short profiles of archetypal users. They will have a name, age, gender, probably a bit of a family or home context and, most importantly a set of motivations and goals for using whatever service is being designed for their use.

Take a moment to think of your Facebook profile.

That has a name, a picture and maybe a bit of a background on where you work and what you do. Most importantly, it also includes your status and wall posts which are likely to betray your expectations, frustrations, hopes, motivations and apprehensions about your daily life.

A persona is basically a bit like a Facebook profile for a representative, typical user, where all the status updates relate to the context within which the new website, service or product, is going to be used (e.g. their job, their home, their plans to book a holiday, etc).

What are they not?

Ah, now we get to the interesting stuff. We’ve come across lots of myths and misunderstandings about personas over the years – all of them reasonable assumptions but sadly missing the point (and thus the value) of personas.

The most important thing to state when discussing what personas are not is that personas are not made up.

For a persona to have any value in influencing and guiding the design process it is imperative that they are based on data collected about real users.

This data can come from a number of sources –  interviews and observation sessions with existing and potential users or even carefully crafted surveys, user testing sessions and also the experience that project stakeholders have had with customers.

These experiences of real users are collected and collated with the intention of finding patterns of behaviour and/or motivation that can start to describe different “types” of user. Each type of user then goes on to form a persona – an entity with a name, age, context and the behaviours and motivations that define that particular user type.

Another common misconception that needs addressing is that personas are not market segments with a name and photo.

Market segmentation is a valuable tool within the marketing profession for categorising a market by demographic and forecasting likelihood to buy and, therefore, the commercial viability of a proposed product or service.

Personas, on the other hand, help inform what it is that this proposed product or service should actually do, based on the recognised needs of the user.

Let’s try and put this within a real-world context: Imagine a high street store has conducted some research and discovered 30 million people in the UK might like a new mobile app to help them with food shopping.

You are no doubt encouraged that your idea is so popular however without knowing the context within which your potential users go about buying their food, how they struggle to buy fresh food and what problems they need a mobile app to solve. There is little to inform the design of this service to ensure that it is relevant to the users’ needs and, therefore, stands a chance of being adopted by those millions of eager beavers.

What are personas used for?

In essence, personas are used as a key focus in user centred design. But, in reality, they can go a lot deeper than that.

Involving key project stakeholders in the development of personas can have a hugely positive impact on the involvement of those stakeholders within the project. Helping to write the personas gives them an opportunity to share their experience and expertise and become engaged in the design process in a way that they can easily relate to. This is particularly beneficial when the project in question is internal to an organisation – for example the development of a company intranet – as engaging users throughout the design process can aid adoption of the new system by producing advocates for it within the target user group.

However, perhaps the most frequent use for personas is to help guide the design of a product, service or solution, keeping the needs of the target users at the forefront, and not the ideas of the designers and developers.

Having design teams refer to the needs, motivations, behaviours and constraints of agreed personas help ensure that content, interfaces and the like are designed to satisfy those users’ requirements. Moreover, referring to these personas can help overcome disputes within design teams or between stakeholders over particular design decisions. Rather than the dispute being focussed on the wants and the needs of the stakeholders, the design decision can be assessed against the needs of the persona – would making this decision be of benefit or detriment to “Thom”? If it makes Thom’s life better, keep it in. If it’s going to make Thom’s life harder it gets dropped.

What if we have more than one type of user?

When there are more than one types of user identified multiple personas can be created. Depending on what is suitable for the solution being designed (and on available resources), you might find that it is appropriate to build several interfaces or pieces of content, etc, to address the different needs of each persona.

Ideally, however, a primary persona would be selected – whose wants and needs are shared by most or all personas and are therefore always satisfied first. Just because Thom prefers fresh food, doesn’t mean you can’t have a dried food section to meet Sally’s needs, but the fresh food emphasis will be stronger.

Why you should use personas

Personas are fantastic tools to be used in the user centred design of any product, solution or service. They help to create cohesion within and between teams of designers and stakeholders and they help to keep the focus of design on the needs of real users which, in turn, helps to ensure that the product you design is relevant and has the best chance of success.

The key to successful personas really comes down to two things – for personas to be successful:

1. They need to be realistic – based on real data from real users, gathered through real encounters such as interviews and observations.
2. Everyone in the project team – from designers, users, managers – needs to buy in to them. Having their involvement in the creation of the persona from the beginning will really help with this and can go a long way to raising the credibility of the personas and its value as a reference point for making difficult business and design decisions.

Want to know more?

If you’d like to know more about personas (and believe me, there is plenty more to be learned!) then why not get in touch with us – drop us an email (more@bunnyfoot.com), a tweet (@bunnyfootsays), whatever suits you.

Or, if you’d like to do some lunchtime reading, here are some useful articles from the web:

http://www.uie.com/articles/personas/
http://www.steptwo.com.au/papers/kmc_personas/index.html
http://www.cooper.com/journal/2002/03/reconciling_market_segments_an.html

Finally, for those of you who really want to get further into this topic, here are some great books for you to get your hands on:

About Face: The essentials of user interface design
Alan Cooper 

The inmates are running the asylum
Alan Cooper

The User Is Always Right
A Practical Guide to Creating and Using Personas for the Web
by Steve Mulder with Ziv Yaar

We developed ‘mass user testing’ in response to the real world needs of commercial clients and to combat the deficiencies inherent in the most widely used traditional usability testing methods (we have actually been doing this for about 4 years but formalised it last year).

The key to mass user testing is using large numbers of people rapidly and cost effectively – this is achieved through recruiting people ‘off street’ with the lure of some cash (or other incentive – we are quite creative in this regard) for about 15 minutes of their time.

Traditional versus mass user testing:

Traditional testing methods have test sessions that last about an hour (sometimes more), and typically 5 or 6 pre-recruited (and therefore expensive) people are tested per day.

Performed well this approach delivers good qualitative results – for instance it will identify usability issues in a website process and provide recommendations for improvements – but it fails to deliver valid metrics (unless costs go sky high), and is often bloated in terms of the value that a test on any particular person delivers (e.g. the main value is often in the first 15 minutes or so, or clients may want to focus just on key discrete design changes, such as homepage or landing page modifications).

Mass user testing by comparison delivers:

• testing with 5 or 6  times as many people (30 per day) for equivalent costs – sessions are shorter and participants are recruited ‘off street’ (and therefore cheaper)
• valid behavioural observations – web site visits are often less than 15 minutes in real life
  valid performance metrics – e.g. statistically robust time to complete task, number of errors made etc only with enough participants can this be robust and built into a performance measurement framework
• valid eyetracking metrics – e.g. statistically robust figures for how many people looked at a feature (e.g. advert or call to action), how long on average it was looked at, how many times it was looked at etc.
• valid eyetracking visualisations – e.g. hotspot figures, average visual journeys through a page etc – these are used to drive optimisation above and beyond usability improvements (particularly effective when used with other big number techniques such as multivatiate testing)
• the ideal testing environment for A/B testing of designs – e.g. 30 people see design A and a different 30 see design B – this eliminates order biases but has sufficient numbers to generate valid comparisons of results

A specialist test facility with specialised software and process

Bunnyfoot recently opened a new office specifically to deliver mass user testing. Reading, in the Thames Valley, was chosen specifically to take advantage of demographics representative of most clients’ needs (e.g. high A, B, C1, C2 without London bias) and a high footfall for recruiting test participants ‘off street’.

The testing was developed originally to meet the demands of Bunnyfoot’s clients for quantitative results and insights from eyetracking testing of print adverts, packaging, e-mail and direct mail. This worked so well that it was opened out to Bunnyfoot’s usability clients too, where it has proved highly effective at supporting rapid development cycles and part of a user-centred-design process. Qualitative and quantitative results are returned within a day and the continual measurement allows design effectiveness to be tracked throughout the process, and/or competing designs to be objectively judged against each other.

An innovative ‘omnibus’ service provision reduces the barriers to testing

To further refine the service Bunnyfoot run an omnibus testing service. Each and every week a minimum of 120 people are tested using websites and other media (print ads, TV ads etc) from a variety of industry sectors.

Clients can pay (either one-off or via a subscription service) to have their materials included in the ongoing testing. This proves particularly cost effective if just the effects of a single page (such as a home page or landing page) or small modifications to existing pages need to be tested – the tests for these often only take a minute so it would be highly wasteful to pay for a full blown longer test.

Mass user testing is the golden bullet to all testing? – well no …

Mass user testing is really useful but is not the golden bullet to all types of customer testing – it is particularly suited to B2C offerings with a wide customer base in terms of demographics (strict customer profiles can often be recruited off-street but it takes more manpower, costs more and takes longer – slightly defeating the whole point).

It is excellent as an extra tool in the whole spectrum of user-centred-design testing activities to be deployed when most appropriate and often in conjunction simultaneously with more traditional testing methodologies.

I bet without too much thought you will have a use for it.

The new standard is called the Web Content Accessibility Guidelines 2.0 (or WCAG 2.0), it replaces WCAG 1.0 – the previously recognised global standard which had been in place since 1997 (that’s eons in ‘Internet time’)

What does the new accessibility standard mean for you?

If your site is already accessible – not much to worry about

If you have a site that is already accessible according to the previous guidelines (see Accessibility Conformance Explained for an explanation) then you probably don’t have much to worry about – in the short term you might want to look at the new standard (or get advice from specialists) - and over the coming months look at transitioning what you say about your site’s accessibiltiy in terms of the new standard.

If you are commissioning a new site or site redesign – refer to WCAG 2.0 in your requirements

In all your documentation make sure you update any sections on accessibiltiy conformance or standards and compliance (these should be in there! – see PAS 78 for more advice on what should go in tender documents) to reflect WCAG 2.0 – but any decent agency should know the implication of this – if they don’t then perhaps they aren’t one – so push them on the point.

What accessibility conformance level should you go for?

This is a little tricky at the moment – the safe minimum is single A conformance to WCAG 2.0 – but you might want to specify double A (with a few caveats depending on specific conditions) – I’ll write an update post on this in the near future.

** STOP PRESS ** January 2009: for UK public sector the current advice from the Central Office of Information (COI) is to specify ‘the WCAG 2.0 equivalent to WCAG 1.0 double A’ – they are looking into more specific recommendations in the coming months.

If you are getting an accessibility audit – audit against WCAG 2.0

If you don’t know how accessible your site is at the moment you should get it audited, and if you are going to get it audited then it should be against the new and current standard.

At Bunnyfoot we currently (in fact have been for the last 6 months for some clients, even before WCAG was formally published) give accessibility audit and accessibility user testing results referenced against both WCAG 1.0 and WCAG 2.0 – this makes it easier for clients to understand the new standard and benchmark against where they were before. We will continue to do this for the next 6 months at least.

Well there is a little bit more to it than that (not much though):

The infrared diodes shine light on the person in front of the eyetracker (it’s 14 times less strong than that emitted from a TV remote – so doesn’t burn their eyes out).

A high resolution digital camera tuned to the infrared spectrum picks up the reflections from their eyes (and some other things too, but lets not worry about that)!

Because the angle of the diodes in relation to the camera is known, the computer looks for the dark spot near to the reflections from the infrared and identifies this as the pupil.

A clever algorithm then works out where the pupils are pointing – and so where the person is looking – it does this between 50 and 120 times per second – depending on the model of eyetracker being used.

The computer then stores the horizontal and vertical coordinates of where each eye was looking and gives it a time stamp. It also stores a picture of whatever was on screen at the time.

Simple – in fact so simple you can find instructions about how to build your own eyetracker on the net (I must admit I didn’t really understand these though)!

It is not just about brand awareness or brand recall anymore, the new era of digital innovation provides us with an array of rich media to communicate with the increasingly cynical consumer. Games offer a huge untapped market with a broader profile than typically assumed. 59% of the UK population (26.5million) are gamers and 45% of those are women! Playing games is not just a nerdy boy thing anymore.

The gaming environment offers a unique and broad advertising medium that offers the opportunity to develop tools that add value to the gaming experience and also new creative opportunity. Moreover, it offers the ability to develop bespoke ad campaigns to specific sectors.

The key benefit to a gaming environment is the ability to capture consumer attention. Instead of a diluted environment where consumers are bombarded with irrelevant advertising, games provide a virtual environment where campaigns can be contextually significant, relevant to game content and consequently impacts greater on the player.

Capturing attention, and providing a standardised metric of attention levels has been missing from the industry, until now. What is currently relied upon is number of impressions, CPM’s, play hours and the standard demographics seen across all types of media. This data provides a good starting place, but a more intelligent validation of advertising investment is required.

Shared environments such as games are ‘seen’ in different ways; orientations, spatial contrasts, speed of interaction all result in different individual experiences. Understanding visual attention, and emotive engagement, at a much deeper level of interaction and on a more individualistic basis reveals honest and reliable consumer responses; it elicits consumer feedback that eliminates participant’s often confounding rationale and appeals to a more useful semi-conscious process that explores visceral reactions to game artifacts.

Through our knowledge, experience and original research we have developed an emotive engagement model that can measure, quantify and predict consumer response to advertising in games. In addition to existing measurements being used, we are able to show how consumers respond to brands and in doing so can validate:

• brand persuasiveness
• brand favourability
• brand impact
• brand loyalty
• overall brand effectiveness.

This provides an independent measure on the value of advertising.

Our methodology has grown from our experience with leading brands in the FMCG industry and digital eCommerce. From outdoor advertising, magazine and website advertising through to print, websites, packaging and TV product placement.

The versatility of the methodology means it can be applied to a variety of advertising medium, such as static or dynamic advertising or product placement in console/PC games, mobile and web, and also TV.

At Bunnyfoot we design bespoke research that delivers both quantitative and qualitative results on your advertising campaign. During any stage of the game development cycle, we can apply our research methodology to understand performance and engagement levels of the ad/product placement. At early concept stage, or green light builds, we work closely with the developers to help understand and optimise strategic ad placement. Towards the latter stages of part level builds right up to beta, we can help define, quantify and predict audience engagement and brand behaviour responses.

What does this mean for the industry?

• Game publishers benefit by gaining knowledge to increase the ad value for their franchises going forward
• Game developers gain valuable investment injection
• Advertisers get the chance to develop tools that add value to a gamers experience – brands are much more accepted in games than they are in an interruptive environment – and gain access to targeted market segments
• Agencies gain unique access to accurate measurement tools and new creative opportunity

Microsoft are extremely aware of the need for usability testing, but what was unusual was that they wanted to carry out regional testing – in England (inside and outside London ), Wales and Scotland .

At Bunnyfoot we had been anecdotally aware that there was a difference in internet behaviour of users inside London and those in the South of the country.

Over the years we have carried out numerous user tests in London and noticed that the browsing habits of Londoners and those outside London demonstrate subtle but definite differences.

The Microsoft testing confirmed our anecdotal suspicions by revealing these regional user-interaction trends in a quantifiable way.

Londoners:

• Tend to be the most experienced web users
• On average they complete tasks in a quicker time but were more likely to fail to complete a task

Scottish and Welsh:

• Tend to be less experienced than Londoners and those from the south of England
• On average they take longer to complete tasks but had a higher overall completion rate

Southern English:

• Tended to sit between the London and the Scottish and Welsh extremes
• May be more representative of the UK population as a whole

What does this mean to you?

When you are planning to test your product beware of testing only in London as it may lead to a distorted view of the UK population as a whole. Remember, the population of London is only 11.6% of the total population of the UK.

Londoners are not representative of the UK population so testing only in London can lead to a high number of erroneously reported usability issues leading to increased spending on remedial work and possibly a delay in making the website live. It can also lead to increased costs and even the perceived need for further, potentially unnecessary, rounds of testing.

Accounting for this regional phenomena in your testing may come at a cost both in terms of time and money. Sourcing an available external venue always takes more time than scheduling an internal resource and this can lead to unnecessary days being added to the project length. Serviced viewing facilities are charged at a premium – they exist to make a profit!

In addition, when using a serviced viewing facility in a different geography we often need to be 100% reliant on external recruitment agencies. Again this can lead to increased costs, additional days being added to the project and externally recruited users tend to ‘no-show’ more than internally recruited users.

As an antidote to this Bunnyfoot now has offices in London (Westbourne Park), Sheffield, Reading, and another close to the Scottish Parliament in Holyrood, Edinburgh .

We now have the largest number of user experience labs in the UK , environments designed to provide naturalistic environments for a variety of testing scenarios:

Edinburgh:

• 1 small office / home office set up
• 1 formal office
• 1 sitting room

London:

• 1 sitting room / small office / home office setup

Oxford:

• 1 small office / home office setup
• 1 formal office
• 1 sitting room

Reading:

• 1 small office / home office setup
• 1 formal office

Sheffield:

• 1 small office / home office setup
• 1 formal office

Bunnyfoot are now the best resourced user experience consultancy in the UK. Our resources allow our clients the option to have cost effective access to regional testing without extending project timeline or cost. It’s another Bunnyfoot first!

Summary recommendation:

If your target market covers the whole of the UK do nationwide testing – avoid the risk of regional bias.

• Is the creative going to make consumers look at the brand and the product?
• Which creative will have better brand recall?
• Which creative has the clearest marketing message?
• Under realistic circumstances will the key messaging get through at all? Do they get the message from the limited time exposure?

Historically, answers to the above have been difficult (or impossible) to obtain. Account managers and clients have had to rely on their own judgement and results from generally expensive focus groups. Both give limited insight into what customers will really do.

Focus groups can produce some good insight when used in the right circumstances. They give the consumer a voice, but they don’t give insight into the consumer mind or emotional attachment. Unfortunately what the mind thinks, the voice often doesn’t say. Consumers can’t always tell you how they will behave, instead you need an alternative approach to see how consumers actually behave with your product and/or campaign.

A quick example.

Of the two designs below which would you sign off?

And now? – with looking at the eyetracking heatmaps produced by testing with real consumers (the hotter the spot the more the consumers looked in that area).

Changing where the model looks changes consumer behaviour. In the advert where the model looked at the product (instead of out of the page) consumers spent a proportionately longer time looking at the product name than the strap line. For this ‘cued’ advert subsequent recall tests showed that the product name was also more frequently remembered.

Well yes we know its a trivial example – but you get the picture. Subtle changes can change behaviour substantially – and it is only by measuring the real behaviour that we can tell this.

Eyetracking provides a new and sophisticated form of consumer testing that can easily match the cost of your focus groups and provide insight into what your consumers do, think and recall about your product and campaign.

Test real impact

Give us a single or a range of creatives and we can provide you with key insights into what people really pay attention to. You can use this information to choose between a set of competing designs, or to drive improvements in an iterative fashion.

Our tests will show you the effectiveness of the portrayal of your brand and key messages under realistic conditions – we place the designs in the most appropriate setting and test what people actually do.

* Tip: you should go for a minimum of double-A compliance, find out what this means below

WAI accessibility conformance and priority levels

The World Wide Web Consortium (W3C) is the body responsible for setting Internet and Web standards. The Web Accessibility Initiative (WAI) is part of the W3C and is responsible for promoting and supporting accessibility.

WAI guidelines and priorities

WAI provide documents and resources to assist with improvements in accessibility. The guideline document that applies to the production of web pages is the Web Content Accessibility Guidelines version 1.0 (WCAG 1.0). This comprises 14 guidelines split into 65 separate checkpoints. The checkpoints are classified into 3 priority levels according to their impact on accessibility. WAI describe the priorities as below:

Priority 1
A Web content developer must satisfy this checkpoint. Otherwise, one or more groups will find it impossible to access information in the document. Satisfying this checkpoint is a basic requirement for some groups to be able to use Web documents.

Priority 2
A Web content developer should satisfy this checkpoint. Otherwise, one or more groups will find it difficult to access information in the document. Satisfying this checkpoint will remove significant barriers to accessing Web documents.

Priority 3
A Web content developer may address this checkpoint. Otherwise, one or more groups will find it somewhat difficult to access information in the document. Satisfying this checkpoint will improve access to Web documents.

WAI accessibility conformance levels

Reaching a certain level of accessibility commensurate with the WAI guidelines and priority levels is often referred to in documents, such as policy documents, requirements documents or tender documents. Unfortunately the nomenclature surrounding this can be quite confusing and the incorrect terms are often used. Below are shown the terms as specified by the WAI but expect to see a mix of terms in documents you come across, for example, conformance level WAI-A (single A compliance to the WAI Web Content Accessibility Guidelines version 1.0) will often be referred to as W3C priority 1, or WAI level 1.

Implementing the checkpoints confers different accessibility conformance levels depending on the priorities passed. Achieving a conformance level entitles the site to display a WAI conformance icon.

Checkpoints passed	Conformance level	Entitled to display
All priority 1	Single-A (WAI-A)
All priority 1 and 2	Double-A (WAI-AA)
All checkpoints (1,2 and 3)	Triple-A (WAI-AAA)

A way to think about conformance levels

Imagine a building with an access ramp. The ease with which the ramp facilitates users who depend upon it can be considered analogous to the conformance levels:

A bumpy ramp = WAI-A (all priority 1)
- It’s possible to get in but with effort to get over the bumps

A smooth ramp = WAI-AA (all priority 1 & 2)
- Easier to get in as the bumps have been taken away

A traction ramp = WAI-AAA (all priority 1 & 2 & 3)
- Assisted with getting in