The chances are high that your audience across your various channels is diverse and made up of people with different needs, different values, different constraints, and more. That’s why smart marketers develop marketing personas to help them better cater to the various groups of people that their brand attracts.
However, sometimes one or more small groups is so extreme in their behavior that they skew the overall behavior of your channel audience in a way that’s detrimental. These groups are called outliers and can cause multiple problems, including:
Clouding overall performance data, making it appear that you’re doing much better or much worse than you actually are
Obscuring how your core audience is performing
Concealing trends in your customer data, causing you to miss opportunities or fail to address challenges
Skewing A/B test results and leading you to implement the wrong changes or strategies
Let’s talk about some common outliers B2C marketers might encounter across their marketing channels and how to manage potential issues with them.
Here are some common outliers that companies might encounter across their marketing channels:
These people spend or donate massive amounts—like 100 times your average audience member. For example, for one of the department store clients we work with, they have a group that’s only about 1% of their audience that spends at exceedingly high levels, says Doug Sundahl, Senior Director of Strategic & Analytics Services at Oracle CX Marketing Consulting.
“Most of them spend over $100,000 every month—and some of them spend many multiples of that,” he says. “It’s likely that these individuals represent a B2B audience of interior designers, resellers, or other businesses.”
JT Capps, Director of Strategic & Analytic Services at Oracle CX Marketing Consulting, adds, “These mega-spenders may also be unauthorized resellers of products. Their purchases then show up across online marketplaces or international markets, which is a big problem for fashion and luxury brands in particular.”
We also see the effects of mega-spenders among our travel and hospitality clients, says Jared Polidoro, Head of Website Optimization & Personalization Services at Oracle CX Marketing Consulting. “Group reservations are an example that could contribute to revenue outliers,” he says, “with one visitor booking a block of 10 rooms for a family reunion or school trip when their average visitor is booking one or two rooms maximum.
Mega offline spenders who never open emails
Unfortunately, never-openers are not uncommon. However, never-openers who spend a lot—particularly offline—are rare. At a glance, it can seem like your emails aren’t impacting these subscribers, but it could be that the presence of your email is having a brand impact on them and that your subject lines and preview text are driving them to your offline channels.
These email subscribers open just about everything that you send them, but that behavior often doesn’t translate into them being a valuable customer.
“For some clients, I have seen the top 2-3% of the list open at close to 100%,” says Sundahl “On the flip side, they don’t click as often as others, and there are opportunities to try to influence their behaviors.”
Influencers & detractors
Your influencers rarely purchase, but they’re incredibly engaged with your emails, social posts, web content, or other content, which they often share with their large personal audience. The dark side of influencers is detractors, who are non-purchasers who are highly critical of your brand and seek to pressure or punish your company with their behavior.
Mobile website visitors
Some of these visitors may exhibit strange behavior or convert at a much lower rate than other visitors. Often, poor internet connections for mobile visitors and the resulting slow page load times are the reason, says Polidoro.
Unfortunately, bots are everywhere. On social media, bots may be sharing your posts on certain topics.
“When doing social listening, it is really important to ensure topics are set up to reduce noise that comes from spammy bot accounts,” says Vicky Czarniecki, Senior Strategic Analyst for Social Media Strategy & Analytics Services at Oracle CX Marketing Consulting. “Otherwise, your data set is overrun with noise that doesn’t apply to the goal of the analysis.”
On the web, it is important to keep in mind that some marketing technologies account for bots, says Polidoro. “Some tools filter out their own bots,” he says, “while other tools may not do this.”
For bots on your email list, in addition to opening just about everything, they are also likely to be clicking at a high rate—sometimes at a ridiculously high rate.
“We had a project where we identified a lot of bot email signups on a list and click rates in the 99th-percentile were at about 3,000%,” says Clint Kaiser, Head of Strategic & Analytic Services at Oracle CX Marketing Consulting. “We found over 200 Gmail subscribers that were clicking everything.”
That said, not all click bots are bad. For instance, some companies use email bots to help prevent malicious emails from entering their corporate mail systems. Typically, these bots click but don’t open emails, says Kaiser. We classify click bots as beneficial, malicious, or exploitive, and we recommend different approaches for dealing with each type of click bot.
“I've seen employees of a company who open and click everything skew email tests,” says Kaiser. “If you’re judging the winner of an A/B test based on engagement and employees are on the list, it’s possible that they might change the picture.”
Whatever the behavior, sometimes all it takes is just one person to throw off your metrics in a huge way, says Sundahl.
“Years back I was doing a customer analysis and realized that a single outlier was skewing our data,” he says. “It was the CEO’s mom. She had spent hundreds of thousands of dollars on high-price point items.”
Our consultants see three major ways of handling outliers:
1. Excluding outliers so they don’t cloud the behavior of the rest of your core audience
The most common remedy for outliers is to identify and exclude them from your data. Sometimes you can do this preemptively.
“It's important to think through a campaign from start to finish through the eyes of the visitor and then try to account for negative scenarios or potential conflicts to the campaign logic,” says Polidoro. “Planning ahead to account for common or known outliers can minimize their impact on a campaign.
For the web, especially when running tests, we recommend whitelisting employee IP addresses or diverting them into sandbox environments, so their behavior doesn’t skew results. “This empowers marketing teams to play around within their web campaigns as much as they like without affecting any of the final performance data,” says Polidoro.
Even if you can reduce outliers ahead of time, you’ll often need to filter outliers from your data. Sometimes you need to do that manually, and other times your marketing platforms will be able to help.
“For instance, website testing and personalization solution Oracle Maxymiser allows you to filter out outliers at five levels: 99%, 95%, 90%, 85% and 65%,” says Polidoro. “The 95% level will exclude the top and bottom 2.5% of your data points. Looking at reports with these outliers excluded will let you get a more realistic view of your data that allows you to understand which experience is truly outperforming the other without these external factors contributing.”
For example, our Website Optimization & Personalization Services team worked with an American gifts retailer during the holiday season and helped them filter out customers who would make gift purchases for 10+ recipients. “They surmised these were business entities making bulk purchases for clients and partners,” says Polidoro. “While an important part of the overall business, our client was more focused on the 90%-97% of their overall revenue that comes from buyers with a smaller cart size.”
Our Social Media Strategy & Analytics Services team has assisted multiple customers through crisis situations where some customers have expressed their disagreement with a retailer’s decision by vigorously commenting on their social media posts. “It is frowned upon to turn off comments on owned content, as it can send a message to customers who are upset that a retailer is hiding or potentially dismissing opinions,” says Czarniecki. “The tactic that was used in a recent retailer situation when every single social post was hit hard by negative comments was to continue to allow comments while identifying who the main detractors were. We kept one eye on that detractor group for any escalation, while also keeping an eye on our core audience.”
2. Segment this group out and send them different messaging
While outliers are often bad from a data perspective, sometimes they’re good from a business perspective. For example, mega-spenders, influencers, and even employees are good great segments worth cultivating.
Consider doing direct outreach with surveys and polls to learn more about an outlier audience and what they want and why. Then test offers and incentive strategies to maximize value and loyalty.
“For instance,” Sundahl, “some mega-spenders might be re-sellers looking for items that they can mark up and drop ship, or items with discounts so high they are willing to buy in bulk for resale. You can change your messaging to highlight recently deeply marked down items of which you may have high inventory. Understanding their needs will help you align them to some of your business goals and derive even more value from your relationship with them.”
You may already have a program in place for engaging influencers on social media. Here are some tips for building successful relationships with social media influencers.
3. Diagnosis whether there’s a technical or design flaw in the user experience that’s causing the outliers
It’s also possible that some outliers represent opportunities to improve your customer experience. If a segment is underperforming in a big way, it may be due to a bad experience that you’re delivering, rather than anything intrinsic to the audience itself, says Polidoro.
“For example, is it people using a specific device or browser that are the outlier?” he asks. “Is their behavior truly different or is the experience inconsistent and broken for that segment of visitors?”
Whatever marketing outliers you have in your company’s data, it’s important to properly identify each group and then come up with a plan for how to deal with them—whether you exclude them, build a messaging strategy around how to directly address their needs, or evaluate whether their behavior is a symptom of a poor customer experience.
Need help with your analytics and performance reporting? Oracle CX Marketing Consulting has more than 500 of the leading marketing minds ready to help you to achieve more with the leading marketing cloud, including Strategic & Analytic Services, Website Optimization & Personalization Services, and Social Media Strategy & Analytics Services teams that can help you optimize your digital marketing performance.
For more information about marketing best practices, please visit Oracle CX Marketing.