In many ways, web analytics history is defined by the pre-Google Analytics and post-Google Analytics eras. The introduction of a free and relatively easy-to-use web analytics tool democratized access and enabled a culture shift to take place in the digital marketing industry, in the direction of data-driven decision making. Similarly, the introduction and proliferation of low cost, easy-to-use A/B testing tools like Optimizely and Visual Website Optimizer has further propelled this culture shift, allowing a broader spectrum of companies and individuals to analyze website traffic and experiment with website changes.
"Testing is becoming more mainstream. Years ago it was difficult to get a program implemented, but now the emerging technology has reduced resource constraints and made the barrier of entry to testing really low."
"Historically, we saw a lot of designers focused on usability testing using Crazy Egg. Now, marketers and product people are using the tool to inform their A/B testing. This is partly because of Optimizely and GWO prior to that; everybody is thinking about A/B testing."
"The optimization team is shifting from an isolated marketing group to a centralized function that works with all parts of the organization. There is a clear need to integrate optimization earlier in the design process to test new features and new design directions."
There is a potential downside to the technological ease and popularity of running A/B tests. Companies can begin testing without first designing an optimization strategy or making an effort to understand the barriers to conversion.
"A lot of people have the testing religion. They got some uplift from a test and are now testing like crazy. Most are missing the strategic dimension. Sometimes you have to bring in the bulldozer and redesign the whole user experience and business model. It takes courage and a lot of time and resources before you even know if it worked. Testing can't help with that."
The exploding growth in the use of multiple devices to access the Internet has changed how marketers approach website optimization. According to a comScore study, the percentage of people using more than one platform to access the Internet (versus mobile-only or desktop-only) grew from 48 to 56 percent in 2013. Many CRO programs are starting to include a multi-device strategy that takes into account the differing behavior patterns among mobile, tablet and desktop visitors. Most of the thought leaders we spoke to said businesses would soon move away from responsive design, or one website that functions well on any device, and instead create device-specific experiences through m. and t. websites (e.g. m.domain.com).
"We tend to have a unique testing roadmap for each device--mobile, tablet, desktop. Fifty percent of the work we're doing is directly mobile for some of our clients."
"I think there is going to be a backlash against responsive design--this notion that you can have an experience that spans everything from an old Blackberry to an 80-inch flat-screen TV. In reality it often means just reformatting and stretching the same old content. More thought will be put into creating dedicated experiences that actually strip down the functionality for each specific context."
"People have different needs and desires on a mobile device versus on their desktop. You don't want to just take your website and make a great version of it on your phone."
In the technology space mobile-specific point solutions have emerged, and many traditional web analytics and testing tools have added mobile and tablet capabilities.
Personalization is a form of targeting that dynamically tailors website content to the needs or preferences of each visitor at an individual or near-individual level. Amazon is perhaps the quintessential example of personalization, where a customer's purchase and browsing history inform a customized homepage with product recommendations. Personalization requires multiple data points on each user, which is often only possible with some type of authentication or login.
Personalization is much discussed in the digital marketing industry, but not yet widely adopted. According to a 2014 survey by Ascend2, 77% of marketers do not use personalization on their website at all, and only 5% use it extensively. Of those that do use personalization, the majority are manually creating multiple versions of webpages for different customer segments rather than using marketing technology to automate the generation of personalized content.
"In previous years, mass customization was the Holy Grail in direct mail but it rarely made a huge difference. It led to marginal improvements at best. The complexity was not outweighed by the benefits. Personalization for e-commerce is the same--a hypothesis you should test. Don't assume that it's worth the maintenance cost."
"Personalization will absolutely lead to greater conversion. Do you respond better to things that are relevant and personal to you? Of course. It's not a topic to be disputed. Personalization is complex, but it's a journey. A little more relevant is better. Then you keep moving down that spectrum."
"Our clients barely have enough time to create three banners, much less individualized experiences for each customer. A few of our larger clients are really targeting four or five segments. One-to-one personalization has a ton of potential, though in practice we're not seeing a lot of it yet."
"A/B testing was not a muscle most companies had a few years ago. We've seen that change dramatically over the past few years. We see a similar trend with personalization. It is a powerful value proposition and most companies are in the early stages of building that muscle. We see it as a direction the industry is going."
Marketers are increasingly interested in being able to view a complete picture of each customer by connecting online and offline activity. This requires integration between various data sources and usually some type of registration or authentication to identify website visitors and collect information about them.
"We're going to see more websites enticing visitors to log in with their Facebook, Twitter or Google accounts, just so they can have that unique ID and gradually enrich the information they know about visitors. It's a tradeoff. I'm willing to authenticate myself in exchange for a feature or something I didn't have before."
"I think we're at a turning point in the industry. Web analytics has to evolve to be more powerful, to bring together all these technologies: data management platforms, ad-servers, data providers, as well as a company's own website and apps and all their digital assets."
"We see a lot of need from our customers for importing and exporting data from our solution to their CRM, CMS, etc."
"One of the more productive approaches I have seen to data sharing is creation of a centralized engine for conversion and proactive update of data between systems. This approach seems to be more flexible, cheaper, faster to implement, and more responsive to changes in requirements downstream than creating a centralized store for 'all the data' where various systems query a central source. Speaking of 'all the data', what often is lacking in these integration discussions is a frank vetting of goals, then deciding what data is essential to accomplishing these specific goals. Integrating 'all the data' tends to be a significant waste of resources; better to start with the key optimization drivers for the enterprise and test concepts outward from there."
The goal is not simply to integrate all of the available data, but to enable marketing tools to take action based on the data in an automated fashion so that each customer is exposed to the most relevant content and the customer experience is optimized.
"How do you start to have that 360 view of the customer? That's a buzz word. But how do you start treating them offline based on what you know about them online, and vice versa, and all of it based on what you know others like them do?"
"You're going to see analytics tools become less passive--moving beyond simple reporting to enabling people to take action on the data quickly, such as emailing certain individuals directly from the reporting interface."