If the $42.3 billion spent online this past holiday season has taught retailers anything, it’s that capturing customers—and their dollars—online is crucial.

But online is a big place. And mobile, which can seem like an entirely different universe, looms ever larger. So where to even start if you haven’t yet…started? And who should lead the charge?

The modern day merchant must have an intimate understanding of the importance of online and mobile commerce, access to a vast array of customer data, and a strategy for transforming this analytical data into winning online experiences.

In all cases, the goal is to attract and retain both new and returning customers. Whether online novices or experts, business leaders crave insight on how to accomplish this. The question is: who inside the company can embody these traits and help the CEO rule the roost? That responsibility should belong to the chief marketing officer.

A CMO should be somebody who uniquely understands marketing, merchandising, data, analytics and web design, and who can also maintain a creative, innovative organizational structure. IT tends to lean too heavily toward data for data’s sake, while Sales too often relies on revenue and relationships.

Placing the CMO in charge allows for the best of both worlds. Armed with the science of data analysis and the art of consumer engagement, the CMO is well positioned to emulate merchant princes of old and join the ranks of retail royalty. A good CMO can nurture a culture of testing, measuring and learning instead of depending on guesswork and subjectivity, as well as reach out to those on the front lines of customer interactions to figure out what those customers want. The ambitious CMO knows that their company site must be more engaging than the competitions’, as well as a place that customers trust, valuing the available products, services and information on offer. It also needs to be a reliable gateway to actions that grow sales beyond the initial purchase, such as cross-selling and upselling.

What’s the best way to make all this happen? One word: data.

Data is crucial to online retail. It comes in many different forms, the main type being the individual behaviors of current site visitors: which search term or webpage brought them over, what time of day and day of the week they’re most likely to stop by, what recent purchases they’ve already made onsite, what pages they visit and what product categories most interest them. All this pertinent info helps define what the “best content” is for each specific viewer. Other types include customer relationship management (CRM) data and social media data.

The aspiring CMO must then use this accumulated data to gain perspective on what customers want; analytical optimization and personalization tools will aid in this quest. Segmentation sifts through the data to find discrete groups of people with similar traits and/or interests, who can then be targeted and tested with relevant content based on site activity. Product recommendations and other offers are then provided based on what the various groups are most likely to purchase.

Product information tools give customers a deeper understanding of the product at hand—a 360-degree view of an article of clothing, or a close-up of various types of textured materials. User-generated content, like ratings, reviews or social media feedback, also aids and influences purchasing decisions. The savvy CMO uses all these methods to strike the delicate balance between intuition and analysis.

On the face of it, the idea of a rigorously testing and optimizing your ecommerce web site seems like a no-brainer.

After all, who would object to gathering hard data on exactly how customers move through your site, what happens on specific visits, and what unseen speed bumps and sticking points are hurting sales? Who wouldn’t want definitive answers to such questions?

Fact is, though, in many online marketing organizations, there is often surprisingly sharp resistance to technologies such as multivariate testing.

Sometimes, it’s simply because there’s some confusion about how multivariate testing actually works and what the benefits are. (“You’re going through all this to test the color of the ‘Buy Now” button? Really?”)

But the real problem usually is buried a little deeper. With hard testing comes the threat of upsetting or overturning some cherished beliefs, or disproving what people ‘know’ to be true. Marketing and creative teams have traditionally crafted websites around ‘established principles’ and ‘gut feel’ for what works. (What if that isn’t so? What if we’ve been dead wrong all this time?)

The IT people build the site a certain way because it’s simple and robust makes perfect sense in terms of data flow. (But what if that actually turns customers off? What if that makes shoppers bail out before buying?)  The notion of actually testing ingrained practices is a bit scary.  Who wants to have their worldview changed?

In other cases, there’s some reluctance to admitting there’s even a problem.  “We’re getting X% conversions. That’s the industry norm.  Nobody is complaining. Why bother?”

The best approach to getting buy-in for a testing strategy is to position it as less of a of a threat, and more of a way to ‘refine’ and direct the efforts of your creative, marketing and IT teams.  It’s a way to out-smart, out-maneuver, out-think your competitors; it’s not a way to beat up on what we’ve been doing.

Detail the pain, document the bad

First step:  point out the pain points on your site. Are visitors fleeing as soon as they arrive? Are they actually buying or just poking around?  Do they buy once, and disappear forever?

To make your case, use strong graphics, charts, video clips, whatever it takes. But make sure the visuals are backed up by solid data that will speak to your audience: You bounce rate, cart abandonment rate, search engagement and average order value. Make the pain and problems clear, and the idea of a testing program will be much easier to sell.

Add up the losses

Find a way to dramatize and quantify how much money you’re losing in the current situation.  Then show how much more money the company could be making if you could get hard data on exactly what’s working and what’s not working.

Bring up smart competitors, industry darlings

Peer pressure is another go-to tactic for getting reluctant executives to embrace change. Explain how other respected, successful organizations are leaping ahead by definitively testing what they’re doing. For examples, check business publications, LinkedIn, public case studies, blogs and so forth.

Nobody likes to be left behind — if stakeholders see other businesses are implementing test programs to their benefit, it’s more likely they’ll want to do the same. And if competition’s already ahead of the game, even better for you. (Although the alternative is that they’re not and you can beat them to the chase.)

Speak to IT on their terms

Your development teams may feel that site testing is unnecessary, especially if they’ve already evaluated usability and other qualitative factors. They may consider testing their domain, and not appreciate any input from the marketing department.

Address these concerns by speaking their language and giving them concrete information they can actually use in making development decisions. Don’t just have a verbal discussion — provide all those statistical facts in writing, in terminology they know and use. Point out that shifting the responsibility of site testing to marketing frees up IT time and manpower to work on other projects.

Relate it all to the brand

In a critical way, the e-commerce site is the final touch point, the ultimate ‘proof’ of the brand concept. Does the experience on the website match the brand promise? Does it reinforce and capitalize on all the branding efforts so far?  Is there something we can do better? Something in our web experience that is somehow compromising our brand?  Multivariate testing can quantify this very accurately.

“Just try it”

Finally, propose a trial run — a limited test, shown to a small percentage of page traffic. Be sure to track key data points like visitor stats, cost per conversion, and abandonment rates. Once the results are gathered, share them with various departments throughout the company. As more co-workers find out what’s going on, your potential support base will become even larger.

It doesn’t take a record-breaking holiday shopping season to realize that most online shoppers are vulnerable to the advances of competing online retailers. But when it comes to a decline in return customers, blaming a lack of customer loyalty on the competition is the easy way out. Now is the perfect time to commit to tracking the number of first-time shoppers who actually come back in months to come—not to mention, those who don’t. And from there, put a plan in place for increasing customer loyalty.

There are many reasons customers will only buy from a site once. It could be that you offered them something they couldn’t get anywhere else, but didn’t give them a good reason to return when they were there. It could also be the overall usability of your site, a lack of necessary information, a poor checkout process… The list goes on. The bottom line is that they didn’t find the shopping experience memorable—and you may never see them again.

But fear not, there is a proven method for visitor retention. And it’s called personalization—aka using what you know about your online visitors to create highly individualized experiences for them. Personalization can be simple, based on one or two collected insights, or a highly complex interaction of detailed formulas and algorithms. Either way, it’s better than just ignoring tailored content altogether.

Because many marketers aren’t taking advantage of personalization techniques—or they’re doing them wrong—there is a lot of opportunity for those who are. So, instead of blaming a lack of loyalty on the aggressive competition, get ahead of competing sites using testing and personalization to avoid making these three common mistakes:

Misinformed website updates

Companies often invest tons of time and money into a complete website overhaul each year (or even every few months), only to find that the new site fares no better—or even worse—than the old. Instead of being attracted to the sparkly new changes, consumers often feel alienated by the inconsistent and constantly shifting branding. After all, if they don’t recognize you, how will they know you’re, well, you?

Even more, companies often don’t fix the actual problems; instead, they focus on the superficial elements they perceive to be the problems. Solving a problem requires first knowing what it is. Let the actions of your visitors show you what works and what doesn’t using A/B and multivariate testing. Testing your website elements and pages will not only reveal your problem areas, it will reveal where on your site you’re experience positive visitor behavior. All of this information should then go to informing minor (or major) site updates, which can lead to significant lifts in conversions. But our end goal goes beyond initial conversion—we’re focused on retention.

Neglecting mobile and tablet

It’s especially sad when a brand goes all out updating their website, but doesn’t take mobile or tablet responsiveness into consideration. There’s little point in having a gorgeous website that’s impossible to navigate on a smartphone. Consumers now expect the same look and feel across web, mobile and tablet-specific apps or sites, yet each platform has its own unique capabilities—and opportunities to learn more about your customers. This additional knowledge can then be used to personalize user experiences across channels. And vice versa.

Landing page tunnel vision

Landing page optimization is a go-to tactic for many marketers looking for a quick fix, but even the best landing pages can’t sustain customer loyalty on their own. Focusing on single pages distracts from the larger picture, which should be the 360-degree customer experience. Knowing what individual customers do, what they like, and how they prefer to engage with your brand can help accomplish this. This knowledge can come from multiple data sources, online and off.

Behavioral targeting is a form of personalization that relies on using these user actions and preferences to inform custom experiences for individual visitors as they navigate each and every page of the site—including landing pages.