For small businesses to medium businesses (below enterprise level), Google Analytics (GA) is one of the best “free” tools available to better understand one’s data. It provides a high confidence level, is precise (if configured properly), & can provide numerous relevant insights about one’s data. However, GA has plenty of limitations & tradeoffs.
Google Analytics never tells you, “Why?”
Google Analytics never answers, “Why?” about website users’ behavior. This is by far, the most significant limitation to be aware of. Quantitative data, KPIs, & metrics all have their limits. You still need to talk & listen to your customers (even if you have Google Analytics 360 Suite).
Let’s run through a few of the other main Google Analytics limitations.
Table of Contents
GA is not 100% accurate & it never will be. It’s an analytics tool, not a silver bullet cure-all.
There’s a lag in receiving data. This can vary, but a safe cushion is to avoid making decisions based on data < 48 hours old. Particularly, for a standard GA account exceeding 200,000 sessions per day, both reporting & metrics can be delayed up to 2 days.
Spam is a major issue, potentially polluting your data in GA (and we know how important it is to utilize clean data). Unfortunately, there’s no easy solution. There’s one setting to assist with filtering spam, but one has to select it in their View Settings.
There’s a hierarchy or order to how GA processes data. This applies to areas such as reporting, filters & channels. It impacts one’s reports. (If you want to read further about scope type & metrics vs. measurement).
GA data is organized into 4 scope-types:
- User data (each cookie tied to one’s browser)
- Session data (by default, a visit within 30 minutes)
- Hit data (a single action)
- Product data* (Enhanced E-commerce only & only for Custom Metrics/Dimensions)
- User (or visitor or unique visitor) = the number of unique cookies that have been seen over a specified period of time
- New User = first-time users or first-time visitors
- Returning User = returning visitor so they have initiated at least 1 session before.
- A visit within 30 mins. = 1 Session
- New Sessions = First time visitors (no prior cookie, user has deleted their cookies, or a new device used)
- Returning Session = Google has seen their cookie before
- A single action that's timestamped, organized, & connected together based on the Client-ID or User-ID
- Pageviews, Events, & Transactions are the most common Hits
- Hits can have products associated with them
- Hits can be classified as either interaction or non-interaction events
GA is cookie based tied to one’s browser. Therefore, users clearing out cookies, blocking them, using multiple devices, etc. can all skew one’s data.
For example, a single user can have multiple Sessions if each session occurs outside of 30 mins., they delete their cookies, they use another device, or a combination of these. Thus, the totals for Users, Sessions, & Pageviews in GA, can become skewed very easily.
There’s a limit of 20 goals per View in GA regardless of the type of goal. (There are 2 main types of Goals in GA, a Destination (Page) goal or an Event goal. Duration goals & Pages/Screens per session are the other types).
Only Destination or Page goals can be used with Goal Funnels. Thus, one cannot mix and match between Page & Event Goals for a Goal Funnel with standard GA (with GA 360 one is able to).
Goals, can be underutilized in GA, especially for micro-goals. Understanding, tracking, & measuring both macro & micro goals is essential for any business.
Tracking events (Event tracking) that don’t require a page being loaded or refreshed requires additional configurations. Because activity (hits) are tracked by pageviews or a page reloading/refreshing, one has to either make additional configurations to the GA tracking code, use Google Tag Manager, or rely on other tools to understand events or interactions that don’t require a page being loaded or refreshed.
A few examples that typically require Event tracking rather than a Goal, include: playing a video, downloading a document, clicking on “quick shop”, or adding an item to a wishlist.
- If you don’t know what you’re doing especially in regards to filters & configurations, you can corrupt your data. This is why it’s so important to setup multiple views in GA (each property can have up to 25 views) and understand configurations impact on your data.
However, using only the reports section of Google Analytics will not break your website.
- Your own team member traffic is not excluded…at all. The Google Analytics Opt-out Browser Add-on is an alternative option to filter internal traffic.
Direct traffic is an “unknown” figure. We don’t know why someone directly entered our website into their browser or clicked on a link that wasn’t tagged. This is the primary reason why we use UTM parameters or perform campaign tagging; to better understand direct traffic. You’ll never have 100% certainty with understanding all of your direct traffic, but that doesn’t mean you shouldn’t improve your understanding of it.
UTM parameters “…are an imperfect solution.” While Google’s Campaign URL Builder is an important tool to utilize, there are additional considerations to be aware of regarding campaign tagging.
- The Site Speed report default sampling size is 1% of your site’s visitors (It can be increased to 100% with adjustments to the GA tracking code or tracked using Google Tag Manager). This report relies on a mixed number of devices (unless you segment).
It’s best to think of this Site Speed report helping alert you about potential slowness of your website for users. There are better options for describing why your site or pages are slow. (While I don’t know how much, most sites are leaving $ on the table because of slowness). I highly recommend prioritizing improving your site’s speed.
- While we can access numerous details about visitors, GA doesn’t know the network speed of our visitors. They could be accessing your site via: 3G?, 4G?, LTE?, or even via a slower connection.
Bounce Rate out of the box can be extremely misleading. (It doesn’t factor dwell time & doesn’t understand context). Additionally, if one doesn’t have event tracking properly configured on relevant pages, to account for behavior or events such as playing a video or clicking a download button, these types of visitors can be included in the bounce rate.
Keep in mind, Google Analytics was launched in 2005 (i.e. long before social media). Thus, it lags behind in some key areas regarding social media reporting. For instance, the default Channels don’t account for paid social & there can be inconsistencies with social media in some of the reports.
Search queries from Search Console data
only go back 90 days from today’s date.(*Update: With the new Google Search Console Beta, one’s able to now access search queries up to 16 months prior from today’s date*). It still shocks me in 2018, how few of sites use Search Console. While extremely helpful for technical SEO, it can help better understand user search queries (for example, it can help with understanding some of the “Not provided” keywords) & serves as a security watchdog for your site. And, it’s free.
If your site is growing rapidly, there are some growth pains with GA & Search Console. Namely, if your search query volume exceeds 1,000 queries or landing page URLs, you’ll have limited reporting in GA & Search Console. You can only view up to 1,000 rows of search queries or landing page URLs.
For Segments, the sample size is not always 100% accurate. If one has a large amount of data in their segment, they may experience data sampling, which can skew things. The work around is to use secondary dimensions, table filters, & smaller timeframes.
- “Default reports are unsampled in both Analytics Standard and Analytics 360.”
- If one modifies a default report, such as applying segments, filters, secondary dimensions or creating custom reports, these are all subject to data sampling thresholds.
- Sampling occurs when at or exceeding “...500k sessions at the property level for the date range…” one is using.
- GA 360 Suite (premium) threshold is, “…100M sessions at the view level for the date range…” one is using. Google Analytics 360 Suite also costs around $150,000 per year.
10 million hits per month per property, is the limit for tracking a web property in standard GA.
200,000 hits per user per day & 500 hits per session are additional limits, for both standard GA and GA 360 Suite, utilizing the Universal Analytics (analytics.js) tracking code.
- The vast majority of changes in GA do not apply to prior historical data. One exception is the “Custom Channel Groupings”, which can be retroactively applied to prior data.
GA Funnels only work moving forward. (Also, one cannot use Event goals in Goal Funnels with standard GA). We all know that “funnels” command a lot of attention in 2017, but any knowledgeable optimization expert will examine other reports besides just your GA funnel. (There are a host of other related issues that can impact one’s funnel as well from improper setup to lack of segmentation).
The challenges with funnels in standard GA (with default settings) is that one is really looking at sessions & cookies, not unique users or people. Many people are trying to understand user behavior by comparing & equating it with session behavior, which typically is not one-and-the-same.
- Multi-Channel Funnels reports only work if Goals or E-Commerce Transactions are setup properly beforehand.
Google Analytics API can help with automating reporting. However, some data from the API, such as the Multi-Channel Funnels report, have a limited lookback window (it’s only 30 days for this data). Thus, if one is reviewing reports in the webpage GA interface vs. a spreadsheet or dashboard that’s being fed by the API, there’ll be significant discrepancies.
Just as a website has numerous moving parts, so does Google Analytics. This is not a comprehensive list of limitations/tradeoffs with Google Analytics, but it’s some main ones to evaluate & be aware of.
What frustrates you the most about Google Analytics?
Was this article helpful?