Great article on Web Development
Since the introduction of user experience (UX) design, web development across ecommerce and enterprise web and mobile builds has remained relatively stagnant. Although emerging usability factors still contribute to design decisions, it seems a major refresh, rebuild or re-platforming of a corporate or ecommerce site is an exercise with too many potential pitfalls.
In an age when agile development has become mainstream, isn’t it time we looked at how to redevelop websites through iterative improvements rather than massive upgrades?
We now have data at our fingertips that can inform UX – data that tells us about user behaviour and how it can change over time. Practices like conversion rate optimisation (CRO) and the psychology behind conversion should be driving decisions on what we implement and the continuous improvements we should be making.
A traditional web development approach generally takes the following steps:
- Business and user requirements gathering
- Data analysis
- User testing
- Detailed wireframes and prototyping
- User testing
- Visual design
- Content development
- Content uploading
- QA & Bug fixing
- UAT & Bug fixing
- Go live
The main challenge with this approach is that it takes months to do and you won’t know the result of the refresh until after it’s complete, meaning you’ve sunk all your money on the build without knowing if you’ll reap the benefits. This is incredibly stressful for teams who will try to second-guess whether the decisions they’re making will be of benefit for the end product.
Data removes subjectivity
User behaviour should be the foundation of any web development implementation and the only way to know how a user behaves is to track visitors’ actions by first testing alternatives and gathering data. However, the most common scoping practices in web development still rely on results from a focus group, or a hunch – either what the internal marketing group thinks or what a designer might ‘feel’ is right.
Unfortunately, these practices are not a true reflection of customer behaviour and can be highly subjective. A focus group, for example, might be affected by one or two dominant voices or opinions. A hunch from a marketer or designer might be influenced by their preferences and/or biases. Opinions are important to consider as a guide to design, but they don’t take into account actual behaviour.
Employing CRO methods means testing becomes part of the build phase. CRO helps organisations and their developers see true customer behaviour. The user does not know they are in a test, so they interact with the site as they normally would, and you gather data about their behaviour, which then informs you what works and what doesn’t in the presentation of a product or service. Whereas those in focus groups might tell you what you want to hear, the stats don’t lie on the page.
This data is also crucial to overcome subjectivity of design. Often a developer will build what the client believes will work based on a small sample size of empirical evidence, usually the experience of the most dominant person in the room. CRO removes that subjectivity through tests that can validate or discredit these opinions. By all means use the opinions as suggestions, but then test them to make the process of design statistically led, rather than be subjectively led.
Testing reduces risk in changes
When UK financial institution Barclays switched its stockbroking service over to a DIY platform over a weekend, it did not anticipate the intensity of the backlash from customers. What happened? The new service displayed information differently and made other unwarranted changes to functionality. Customers couldn’t access their accounts and couldn’t locate transaction details.
Disaster could have been prevented by implementing data-led UX. If Barclays had instead tested potential changes incrementally instead of implementing a wholesale change over a short period, it would have been able to identify and fix potential issues during testing. Using a CRO methodology means organisations can test first, then implement, instead of investing millions of dollars in a 12-24-month rebuild they don’t know will deliver desirable results.
CRO gives brands the ability to see the impact of each test on revenue within weeks. Businesses can see if a change to a template will better engage a customer or have a negative impact on their experience and thus the brand’s revenue.
Don’t play the long game
Considering the average website rebuild takes 12-24 months and has a sizeable risk attached to it, the hyper-agility of CRO and its data-led certainty is the obvious antidote. A/B testing is not new, but the process of attaining live customer feedback and then using the data in decisions on the structure, message and presentation of a product or service experience is sadly underused.
Shorter sprints mean businesses can refresh their site as they go and learn through the process, determining what works before committing to implementation. It’s a faster, less risky way to build or rebuild, and often pays for itself.
Additionally, these benefits are cumulative. As you test and prove the design changes work via increased conversions, the business has a corresponding uplift in revenue that then pays for the upgrade of the site. This happens over and over, giving you iterative, cumulative benefits that traditional web development can’t.
By contrast, most programs of work begin with long strategy planning and research phases to give the business confidence that they have included customer feedback, listened to all stakeholders and taken into account analytics. It can take months to collect and collate data, derive insights and develop a strategy, let alone see a return, whereas CRO testing can tell you whether or not something works in a matter of weeks.
Secondly, the iterative nature of CRO means you can implement changes to address relevant customer behaviour. When we use analytics data to understand behaviour, we look at past actions to help us understand future actions. Social context has a strong influence on our decisions, however, so using past data to predict the future is flawed because the social context has changed. CRO allows brands to test and learn on the go, rather than interpreting behaviour from 6-12 months back. Fast data—the ability to test, learn and refine—makes CRO contextually powerful.
Web development methodology has remained fairly static since the establishment of UX, so it’s hard to see another way of doing things. But when you pick apart what makes traditional development risky and cumbersome, it’s clear that the hyper-agile, data-led principles of CRO can solve a lot of the problems.
Never underestimate the ability to run a test, see customer behaviour and pivot if necessary—it will soon become the primary way websites ensure they deliver the right outcomes for their organisations.
by Nima Yassini on Mumbrella