E-Commerce Analytics and VPNs: What Businesses Should Know

Here's the catch. Today, consumers increasingly surf the Web behind a virtual private network, for privacy reasons, security when using public Wi-Fi, or just to find a different location that offers a better price on a sale. By some estimates, more than 30% of the world's Internet users are regular users of a slice. That means if a piece of "visitors from Germany" appears in an analytics dashboard, it could be a visitor from Brazil or the Philippines. 

It's a problem for a business that's trying to understand where its customers are truly spending time, and it's becoming an even bigger concern as more customers install quality VPN app options, like that from VeePN, onto their phones and laptops without even thinking twice. When the information that feeds them is off, it can cause everything from pricing and currency defaults to shipping estimates, even fraud checks to go awry. When analytics and VPN apps usage intersect like this, ‘where is my customer, really’ becomes a darned difficult question to answer.

Online retail sales reached the $6 trillion mark globally during 2024, and this figure continues to rise each year. With every dollar spent there's a trail of numbers, such as what kind of device was used, how long did they spend browsing, and what page did they leave from? Retailers use this data to help them optimize their website, from the layout of the home page, to the exact time they send an email. The fact that analytics is one of the most funded areas in the retail technology segment isn't surprising.

Cart abandonment is estimated to cost about $18 billion in lost sales a year for online businesses. That's one metric and that is the amount of money. It depicts the rationale behind monitoring behaviour in the first place. If there is no data, a store owner is more like just guessing. It will help them identify their losing step and offer solutions to correct those sales errors.

The overwhelming majority of platforms are measuring pretty much the same set of things, regardless of the stark differences between dashboard designs. Let's get the overview:

• The average time spent on the page and the bounce rate.Average time on page and bounce rate.
• Keywords: cookies, user behavior, search, social, direct, referral
• How many of the products or categories converted
• The type of device or browser used.Type of device or type of browser.
• Geographic location of visitors
• The location at which a cart is abandoned

The final one on the list is location, which we took two sections to dissect.

There is also another level – not geographic, but consensual. Today's shoppers know more than ever that their actions are monitored, timed and recorded somewhere. Some embrace it because it results in a lower amount of irrelevant ads and smooth checkout processes. Others reject purpose, knowing that cookies are being removed, trackers are blocked, and traffic is channeled through a VPN, all in the name of preventing a store from creating a profile without their knowledge.

“When we began to ask what information we needed to know, rather than what information we could capture, we began to see a change,” said one small online retailer in an industry roundup from last year. It's in the difference between data a business needs and that which it can gather.

All this is not a sign of the death of analytics, or a justification for businesses to toss their dashboards out the window. It implies there is a need for some background to the numbers before anyone takes action. The sudden geographical anomaly could be a VPN, not a new market opening up over night. The lower personalisation accuracy could be due to privacy tools rather than a poorly targeted campaign.

Here are some tweaks that can improve it:

• Just because you experience abrupt and unexpected location changes, doesn't mean it reflects data quality, rather treat it as a data quality flag.
• Rely on first-party data over inferred signals, such as purchase history and account information.
• Develop price and fraud policies without breaking as soon as the location is unusual.Create pricing and fraud rules that do not fall apart when location is unusual.
• Ask customers directly (preference centres) rather than just assume based on their behaviour. 

Just also try to take the perspective of the customer on the screen for a second. A late night shopper during an online comparison in the car is not trying to destroy the store's dashboard. They are being prudent, like locking the front door because it's just a habit, not because of fear.

Some online shoppers just use an extension for no particular reason; it could be one of the many free options on the Chrome Web Store. Companies that factor that into their analytics design will have more “clean” information, fewer customer frustrations with a needless fraud check stuck in the middle of the line, and more true insights.

GDPR, CCPA, and similar rules have made IP addresses a sensitive topic. Many analytics platforms now anonymize IPs by default. VPN usage adds another layer. When a user connects through a VPN, their IP is already a mask. That can be good for privacy, but it makes compliance harder. You might think you are not collecting personal data because the IP is fake. But the rest of the session could still contain emails, order details, or device fingerprints.

Businesses should update their data retention policies to account for VPN traffic. If you cannot reliably link a session to a region, do not force it. Store the raw data for a short time, flag it as “low confidence,” and move on. Overcooking the data just creates false segments.

Online analytics is not a thing of the past – and that's a good thing. It remains the best barometer of what actually works and does not. However, the Internet that it users access has evolved, surreptitiously and permanently, and the data needs to be interpreted accordingly. In the end, a touch of skepticism about any single metric, more trust in first-party data and a true appreciation of customer privacy will go much further than a push for perfectly clean numbers that in reality do not exist.