Why Aren't My Google Analytics Numbers Accurate?

If your Google Analytics (GA) numbers aren’t accurate, it might not be your fault. GA is a great tool for measuring your online performance, but some of these ostensibly straightforward numbers can be misleading. We’re going to cover the four most common metrics that can lead to questionable conclusions, how to avoid some pitfalls, and how to get to the metrics you actually want. If your experience with GA is like most, with these alternative data points, you’ll likely be able to solve a few long-standing mysteries, uncover some valuable hidden data, and make better informed decisions.


The Problem: Time on Page & Session Duration

This is one of the most misunderstood metrics in Analytics because Google doesn’t mention the one major caveat—Google Analytics can’t measure the time spent on the final page of a visit. Google always gives the final page in a session a duration of 0. If someone spends 5 seconds on the homepage and clicks through to your blog and reads a post for 10 minutes, the session duration is only 5 seconds because the blog page has 0 seconds for time on page. Pretty major discrepancy, right?

The Fix:

  • Google Tag Manager is the tool for this job since it allows you to set timers to see how long users spend on all of your pages. Tag manager is a topic for another post, but if you want to learn more, you can start here.

  • Setting up several timers at intervals of 10 seconds, 30 seconds, 1 minute, 3 minutes, and 5 minutes will give you a good starting point to start accurately measuring page and session durations.

  • It is important to note that Analytics has an option for a duration goal, but this will also not work on the last page visits. Any time related goals should be based on events setup in Tag Manager.


The Problem: Bounce Rate

High bounce rates are bad and low bounce rates are good, right? Well, not so fast. Bounces occur when a user visits a page then leaves without clicking or interacting with the page. High bounce rates are often seen as an indicator of a poorly designed page, but there are numerous instances when this isn’t exactly the case. Pages with high bounce rates typically either appeal to a small segment of users, or simply don’t offer interaction points for users. If your page is close to the standard bounce rate range of 41-55%, there’s no need to worry.

The Fix

  • For pages that you expect users to interact with or click on, you should make sure that there are plenty of opportunities for that, not just descriptive text and a couple sets of menus. In addition to links, there should be a clear call to action to convince users to interact with your page.
  • Since every site has unique purpose and design, when evaluating your bounce rates, focus on internal benchmarks. How does each page compare to your homepage and landing pages which should have the best interactions (low bounce rate)? And when making improvements for interactions, focus on improving each page’s progress over time.
  • If you have a page that is meant to only be read, such as a blog or product specifications, there aren’t many places to click and a high bounce rate should be expected.
  • Not all pages are for everyone, and some pages don’t need to have high interactions. Pages for specific products, job postings, specifications, and news posts, etc. are expected to have bounce rates higher than your homepage.

The Problem: Direct Traffic

Direct Traffic is typically assumed to represent visits where the user typed in the URL directly— but this can be a misleading assumption. GA is able to trace where visits come from and group them into different channel buckets, but what few people realize is that all of the untraceable traffic goes into the Direct Traffic Channel Bucket. There are two types of traffic that wind up here; the first is the correctly classified users who visit your site by typing in a URL directly or using a bookmark. The other type of traffic is mislabeled as direct, as it lacks the proper tracking to place it into its correct channel. This mislabeled traffic is called Dark Traffic, and it causes traffic from sources like social media or email campaigns to be misclassified as Direct Traffic. Dark Traffic is caused by either a traffic source that isn’t automatically tracked by Analytics (links in emails, PDFs or software, to name a few) or the tracking data gets stripped—which can happen with manually sharing URLs, moving from secure (https) to a non-secure (http) site, a URL redirect or using URL shorteners, such as bitly.

The Fix:

  • For sources that aren’t automatically tracked by Analytics, adding a tracking code to the end of your URLs will let Analytics track where your traffic is actually coming from. Google’s Campaign URL Builder can generate these codes for you.
  • To help curb manual sharing, share buttons are a good way to limit copying and pasting of URLs while keeping any tracking codes in the shared URL.
  • Minimize use of URL shorteners or re-directs as much as possible.
  • For site promotions outside of the digital world, creating unique landing pages for each campaign does a reasonably good job of ferreting out this type of traffic. This will require some extra setup on the website, but the benefit of a customized experience for visitors will be worth it.

The Problem: Sampled Data

If you run a report with more than 500,000 sessions, Analytics will automatically sample your data. Sampling uses a portion of your website’s traffic to report what happened for all of the traffic. This helps speed up reporting for large sites or long date ranges, but can also be the reason why traffic on different reports isn’t adding up. Anytime sampled data is used, there will be a message in the upper right corner to tell you how much of a sample was used.


The Fix:

  • The drop down menu to the right of the sampling message allows you to choose greater precision, which will give you more accurate data if you’re willing to wait a couple more seconds.
  • Another option is to choose a shorter date range, which lessens the amount of sessions Analytics needs to sort through.

Consider Yourself “In the Know”

With a better understanding of some of Analytics metrics, along with these take-aways, you’ll be able to report more accurate data.

If you have questions about how to more effectively implement or interpret your analytics, we’re ready to help. Just contact us. Or, click around our site to see all of the other ways we help our clients outsmart, outmessage and outmarket larger competitors, to steal their most profitable customers.