Article

Negative Events + Other Ways To Figure Out What’s Wrong

Advertising Analytics Web Development

June 25, 2023

I want to focus on strategies to uncover the hidden issues, pitfalls and blockers that may be preventing your visitors from converting at the rate you’d like. This is in response to a question someone recently asked me, “How do you determine what to optimize and when there’s an actual issue with your site versus an issue with ad platforms or traffic quality?”

I believe the answer comes down to a few core tasks, each of which I’ll cover below:

Building an “early detection” system to flag potential blockers
Uncovering meaningful performance differences across groups of pages
Identifying specific on-page issues that harm task completion rates

Building An Early Detection System:

The last thing any marketer or operator wants to hear is that s/he needs to manage (and pay for) yet another platform/reporting system/piece of marketing technology. And the good news is you don’t have to! Instead, we’re going to turn to a new, old friend: Google Analytics 4.

I’ve long been an advocate for setting up Negative Events in Google Analytics.

Negative Events are exactly what you think: events that fire when a specific problem/issue occurs. By proactively listening for these issues, you can pinpoint & address them well before they rise to the “this is harming our business” level of impact.

This is another area where GA4’s event-driven model shines, as you can add multiple parameters to each negative event – giving you even more insight into what’s causing an issue for your website’s users.

A few examples of negative events that we use:

Internal link to 404 Page – one of the easiest ways to tank your conversion rate is to (inadvertently) have a CTA lead users to a 404. It’s just a bad experience. Fortunately, you can create a trigger that fires when a user clicks an internal link AND ends up on a 404. Include the parameter of the url that was clicked, the link text, and the page where the link was clicked. Voila! If you start seeing that event firing, you’ll know exactly where to find the broken link + address it.
Out of Stock (OOS) Products – OOS is the silent killer of ecommerce conversion rates – but keeping track of stock, especially on sites with hundreds of thousands of SKUs, can be near-impossible. This is another area where GA4’s event-driven model + product schema shine. With a little work in Google Tag Manager (you’ll need to set up a DOM element variable in GTM to capture the “out of stock” text from your site – CSS selector method generally works well, but will depend on exactly how your site displays the OOS message), plus GA4’s product schema (both Product Title + Variant ID are included in the enhanced schema), you can fire events each time a view takes place of a given OOS variant.

If you want to take this event further, you can create a report that shows OOS rates for each SKU. There are multiple methods you can use for this, including: (a) including a custom parameter with the total number of variants in a SKU or (b) firing an “in stock” event (similar to OOS above). With that data, a simple custom metric will enable you to view OOS Rates by SKU ([Unique Variants w/ OOS Event] / ([Unique Variants with IS Event] + [Unique Variants w/ OOS Event]).

You can also pull a handy custom segment in the “traffic acquisition” section of GA4, showing exactly what sources are driving traffic to OOS variants – and either (a) pause those ads/campaigns OR (b) add a data capture point, so you can recover those sales as soon as your product is back in stock.

Form Field Progression Issue – track which fields in a form users aren’t completing during abandonment. This data can provide insight into whether a form is too long, contains redundant information, or whether it is better configured as a multi-stage form. You can track this in multiple ways, but the easiest is including multiple events for “form_data”.

If that’s a bit too much work, GA4 tracks “form_start” and “form_submit” events by default in enhanced measurement – so a simple calculated metric of “form_submit” / “form_start” will give you a form-by-form.
Form Field Validation Failure – is there anything more annoying than entering your phone number into a form, only for the form to spit back a rejection because your area code wasn’t in a parenthetical? Fire an event for “form_submit_failure” with a parameter for “reason” = [validation text]
Site Search – site search is both a goldmine for UI/UX data and a point of friction for users. Simply tracking the text of the search can yield a wealth of insight – along with the specific page(s) where those searches take place.
Low Scroll Depth – Are most users simply not scrolling through your content? You can fire a simple report that shows the percentage of users who get beyond the initial viewport / initial 25% scroll.
Help/Support Contact – If someone’s looking for help/support, there might be an issue. If those people are coming in from paid search – perhaps some segmentation in your brand campaign might be in order.
Sessions w/o Conversion – finally, I like to create reports for pages with a minimum number of sessions (say, 500) with less than conversion. This isn’t a negative event per se (more like a negative report), but one which might indicate an issue with on-page tracking OR an issue with the page itself.
Quiz Failure – similar to a form abandon, I like to track quiz progression + failure rates. If you track this by screen/question, you can get a sense of if your quiz is too long, or if certain questions aren’t making sense to your audience.
Video Issues – track to see if users are having issues with your video content buffering or simply not displaying.

These are just some examples to get you started + thinking – you should absolutely customize them to your organization’s unique needs.

Identifying Patterns Across Pages:

Negative conversion goals, on their own, are powerful. But they’re even more powerful when paired with a more in-depth understanding of your site. To do this, I recommend creating a Custom Content Group for each section of your website OR page template/page type. For instance:

Home Page
Product Category Page
Product Category (i.e. all PDPs for a specific category, like “shirts”)
Product Display Page / Service Page
Resources Page
Blog Page
Contact Page
Support / Help Page
Careers Page
Job Listing
Lander

Use this Custom Grouping PLUS your negative conversion events (above) to pinpoint issues that might be structural in nature – for instance, a page layout that doesn’t work for your users or an issue impacting a certain sub-section of your site.

Identifying Specific On-Page Issues:

In some cases, the above two methods will be more than enough for you to pinpoint a specific issue + resolve it. In other cases, the issue might be more subtle or more nefarious – perhaps users aren’t understanding your content, or they’re rage-clicking when a quiz or tool doesn’t work as expected.

To do this, I highly recommend you install Microsoft Clarity. It’s absolutely one of the best heat mapping + session recording/analysis tools out there. In addition to being a great platform, Clarity also has a few other advantages:

It’s 100% free, forever
It’s GDPR + CCPA compliant
It does NOT use sampling
It integrates with GA4

The “Insights” section of clarity provides exceptional reporting on everything you’d want to know, including:

Rage Clicks – are your users rage-clicking on an element of your page? If so, that’s a good indication that something isn’t meeting their needs and expectations. Once you’ve identified a page with a high rage click rate, you can view the session recordings to see exactly what has triggered the issue.
Excessive Scrolling – are users scrolling all over the place during a visit (i.e. up and down and up and down)? That usually indicates an inability to find exactly what your user wants – and presents an opportunity for better information structure.
Quickbacks – does a user click on a link, only to return to the previous page near-instantaneously? If so, there could be a content mismatch or other issue.
Dead Clicks – does a user click on something (a link, a tool, a video, etc.), only for nothing to happen?

Along with example recordings of sessions where each issue took place. This makes addressing those issues a little easier and a little quicker.

Found Problems, Now What?

Uncovering issues impacting your site is the first half of the battle; the second is determining an appropriate response. In some cases (for instance, validation errors or broken links), that’s easy: fix the issue.

But what about for more nuanced or insidious issues, like low page engagement, minimal scrolling, low conversion rates on a specific page type or excessive scrolling? In these cases, my recommended next step is an experiment, designed to target the specific pain point.

Related Insights