
When your Google Ads stop converting or experience a sudden drop in performance, your immediate reaction might be to question your budget or the broader market conditions. The most likely culprit is actually broken conversion tracking. When your measurement setup misfires, Google’s Smart Bidding algorithms begin optimising toward flawed data; this causes your reporting to feed you inaccurate information, meaning every rand you spend actively works against your business goals. Fix the measurement foundation first.
Two audits we ran earlier this winter make the point. Both of these businesses approached us with the absolute conviction that their ad campaigns were failing. One was a prominent education company spending a significant amount of money across both Google and Microsoft Ads. The other was a highly active, lead-driven services business. These companies operated in completely different industries and targeted different demographics, yet they shared the exact same frustration: the ads used to work, and now they don’t.
Neither had an ads problem. Both of these accounts suffered from severe measurement issues.
In the first account, we found conversion actions double-counting in several places while failing to fire entirely on other crucial pages. On their Microsoft Ads account, absolutely nothing was configured correctly; they were running campaigns with zero usable conversion data. The second account presented a different challenge entirely. They had a high volume of valuable leads arriving via phone calls and direct WhatsApp messages. Those critical interactions never made it into the advertising platform’s view of reality, leaving the system completely blind to their most successful channels.
You are no longer manually managing bids on a granular level because the machine learning algorithms handle that process. Those algorithms only know exactly what your conversion tracking setup tells them. Feed the system inaccurate data, and it will confidently spend your entire daily budget chasing low-quality traffic at scale, 24 hours a day.
Profit, not clicks. That’s how we run every account at Tomorrow Labs. You cannot optimise an account for tangible profit if the system cannot see where that profit originates.
You do not need an experienced developer to spot the warning signs of a broken setup. Run through this practical list against your own account to see if things are misaligned:
If you recognise two or more of those symptoms in your own account, you should stop tweaking your ad copy immediately. You are dealing with a severe data problem that requires immediate attention.
Automated bidding strategies like Target CPA, Target ROAS and Maximise Conversions set your auction bids automatically based entirely on the historical conversion data your account reports. This is a fundamental mechanical reality of how Google describes its own system.
Follow the logic of how this mechanism actually plays out in a live account. If your tracking setup consistently over-counts interactions, the algorithm starts believing that cheap, low-intent traffic is highly valuable and aggressively buys more of it. If your tracking under-counts your real leads, the algorithm assumes your highest-converting audiences are failing and stops bidding on them entirely. The system learns the wrong lesson from your flawed data and systematically applies that bad lesson to every single auction you enter.
This dynamic explains why business owners often feel their ads have suddenly degraded in quality. The campaigns did not lose their effectiveness; they are simply executing perfectly against a corrupted definition of success. Lunio’s detailed analysis of Smart Bidding chasing bad data documents the exact same failure patterns we regularly identify in local South African accounts.
While your underlying tracking remains broken, your entire reporting suite is basically fiction. Metrics like cost per lead and return on ad spend become completely untrustworthy. Any strategic business decision you make based on those numbers is essentially a wild guess dressed up in a polished spreadsheet.
This issue is highly prevalent locally because of the specific ways South African consumers prefer to communicate with businesses. We operate in a heavily phone and WhatsApp-driven market. A massive percentage of high-intent enquiries in sectors like property, automotive sales, higher education and B2B services will never interact with a traditional website contact form. If your measurement strategy stops at the form submission button, you are optimising your entire marketing budget based on a tiny fraction of the truth.
We also see massive issues with platform sprawl. Account managers often stack different tracking methods on top of each other over the years. You might find GA4 events imported directly into Google Ads sitting alongside a legacy Universal Analytics tag that was never removed. You might have a WordPress form plugin that changed its post-submission behaviour during a routine software update, subsequently breaking the firing sequence. Every single layer of complexity introduces another opportunity to double-count a user or drop the data entirely. Microsoft Ads often suffers the most in these situations; we regularly audit legacy Bing accounts where absolutely zero conversion tracking has been configured.
Incentives also play a significant role in this problem. Inflated conversion numbers can make monthly agency reports look fantastic to a client who does not know how to verify the data. Agencies are rarely eager to audit the fundamental figures that justify their ongoing monthly retainers.
We rely on a very specific sequence of actions during our tracking and analytics remediation work to ensure everything functions perfectly:
Ask yourself one critical question before skipping this process: can you easily trace this month’s reported platform conversions back to the actual, verified enquiries your sales team received? If you can confidently do that, you are in a good position. If you cannot answer that question affirmatively, you need to start at step one immediately.
Your objective is a tracking system that accurately reflects bottom-line profit; a setup that merely functions on a technical level is not enough.
A primary conversion action should represent the closest measurable digital moment to money physically changing hands. For an e-commerce retailer, that means tracking a completed checkout alongside the exact cart value. For a B2B lead generation company, that means tracking a highly qualified prospect. You ideally want to import backend CRM outcomes to tell Google exactly which specific leads converted into paying customers. When you reach that level of measurement maturity, the algorithm will begin spending your daily budget the exact way your CFO would want it spent.
That is the entire objective of paid media management. We are not looking for higher click volumes or cheaper cost-per-click metrics. We want to generate maximum provable profit for every rand spent; you need to present numbers you can confidently defend in a board meeting.
An audit of a typical account takes days, not months. Remediation depends on the extent of the problems – simple tag fixes are quick, while rebuilding measurement around calls, WhatsApp and offline imports is a larger project. Either way, it is faster and cheaper than another quarter of misfiring ad spend.
Real results improve once Smart Bidding relearns from clean data, which takes a few weeks. Reported results may initially look worse because inflated numbers disappear — that is the measurement becoming honest, not performance declining.
Yes. Compare a month of reported conversions against your actual enquiry records, then test-fire every form and tag. If the numbers reconcile, your tracking is in decent shape. If they do not, get professional help – this is precision work.
Yes. Meta’s delivery system learns from pixel and Conversions API data in the same way Google’s Smart Bidding learns from conversion tracking. Broken signals waste budget on every platform that uses automated bidding.
Sometimes, but you cannot diagnose a demand problem through a broken measurement layer. Fix the conversion tracking first; clean data will then show honestly whether the demand exists.
Jean-Luc Terblanche is the CEO of Tomorrow Labs, a full-service digital agency in Cape Town that builds marketing around one idea: profit, not clicks.
If your gut says your numbers are lying, it’s probably right. Book a conversion tracking audit with Tomorrow Labs and we’ll show you what your account is really doing before you spend another rand.