When 60% of Orders Need Customer Service, Your Ecommerce Data Is Lying to You

A site can be profitable and still be telling leadership the wrong story.

That is the uncomfortable part of ecommerce reporting. The dashboard can show conversions. Orders can keep moving. Revenue can keep everyone calm. And the whole time, the website may be getting credit for work that people inside the business are still doing by hand.

That was the Farmers Depot problem.

We came in through a partner agency to support media buying. Before we got anywhere near a spend plan, we reviewed the site and saw enough friction to stop pretending the issue was just traffic. Important interactions felt awkward, inconsistent, or unfinished. The site looked like ecommerce from a distance. Up close, it did not look like a system that was reliably closing the sale on its own.

The easiest move would have been to take the budget, buy traffic, and let the reporting keep flattering the setup.

We did not think that was defensible.

This reads like growth in a deck and like friction in the funnel.

Then the more serious issue showed up.

Once we got into analytics, the measurement layer itself stopped holding up. GA4 was thin. Other key Google tools were missing. The setup was not giving the business a trustworthy read on what people were actually doing. And then came the fact that changed the conversation: roughly 60 percent of online sales were being processed by customer service through the website.

That is the point where an ecommerce dashboard stops being a clean record of self-serve behavior.

The order may still count.

The revenue may still be real.

But the website did not necessarily do the conversion work by itself.

‍

When the Website Gets Credit for Staff Labour

This is where the dashboard stops measuring reality and starts managing feelings.

If a customer gets stuck, calls support, and a staff member finishes the order inside the same system, the platform will usually log the sale. Revenue still lands. The dashboard still lights up. The business still gets to say the website converted.

That is the convenient fiction.

The sale happened. The site participated. But participation is not the same thing as self-serve success.

That distinction matters more than most teams want to admit. Once staff-assisted ordering becomes a meaningful share of the transaction flow, the conversion story stops describing what the customer journey can do on its own. It starts blending customer intent with human rescue work.

That is not fraud. It is not misconduct. It is what businesses do when the digital experience still needs help and the orders still have to move.

It is also how leadership ends up funding the wrong fix.

If the site is getting unearned credit for conversion, teams start diagnosing an acquisition problem when they are actually looking at a self-serve problem, a measurement problem, and an operational workaround all tangled together.

‍

Revenue Can Cover Up a Weak Self-Serve Experience

Revenue is one of the easiest ways to avoid the harder question.

If money is moving, people assume the site must be working well enough. Maybe it is clumsy. Maybe it needs cleanup. But if orders are happening, surely the next move is traffic.

That assumption protects weak systems.

Farmers Depot was already producing meaningful revenue while the site still showed obvious signs of friction and weak measurement. That is exactly the kind of setup that misleads operators. Revenue can confirm demand. It cannot confirm that the digital path is healthy, that customers are succeeding without intervention, or that the reporting is describing real customer behavior cleanly.

Once the baseline was cleaned up, the picture got less flattering fast. Internal and call-center traffic had been distorting the benchmark view, which meant the business had more confidence in the digital path than the site had actually earned.

That is the trap. The numbers do not have to be fabricated to become strategically misleading.

A dashboard can be numerically accurate enough to pass a meeting and still point leadership at the wrong decision.

‍

Why the Media Conversation Had to Stop

The original ask was media support. The real issue sat earlier in the chain.

Once we understood how much assisted ordering was being folded into the same system, a media-first conversation stopped making sense. Buying more traffic into that setup would not have clarified performance. It would have made the misunderstanding more expensive.

That is the mistake too many businesses make. They use spend to avoid diagnosis because spend feels decisive and diagnosis feels inconvenient.

But more traffic into a distorted funnel does not produce clarity.

It produces a bigger pile of numbers that still cannot tell you what the website actually did.

If the business cannot separate true self-serve conversion from staff-assisted completion, every downstream acquisition discussion gets weaker. Channel tests get weaker. Efficiency reads get weaker. Conversion arguments get weaker. The confidence stays loud. The evidence stays muddy.

That is why the media question had to wait.

‍

What We Needed Before Acquisition Could Mean Anything

At that point, the work stopped being about opinion and started being about visibility.

We needed a baseline that was not polluted by internal activity. We needed cleaner instrumentation. We needed a clearer view of where people were stalling, what they were missing, and how much intervention the business was still supplying behind the scenes.

That meant auditing GA4, establishing an IP-filtered baseline, installing missing Google-suite components, implementing event tracking, documenting UX friction, and building a roadmap for what had to happen next.

Just as important, we needed to trace the journey closely enough to stop letting the loudest guess win. User-path tracking and heat mapping gave us a better view of where customers hesitated, where they dropped, and where internal assumptions about the journey were simply wrong.

That work is less glamorous than campaign launch. It is also the work that keeps a business from paying to amplify a story its own reporting cannot support.

This is the part teams underrate. Data cleanup is not clerical pre-work. In situations like this, it is the thing that makes the rest of the growth conversation honest.

‍

What Changes Once You See the Problem Clearly

If your site depends heavily on staff assistance to close orders, the usual ecommerce questions are too shallow.

Do not just ask whether conversion is up.

Ask what kind of conversion the number is describing.

Do not just ask whether revenue is holding.

Ask how much of that revenue still depends on people inside the business stepping in to finish the job.

Do not just ask whether paid media should scale.

Ask whether the system can tell you what paid media is actually doing after the click.

That is the real lesson here. The danger is not just messy analytics. The danger is that the business starts treating a blended operating model like proof of a strong self-serve experience.

Those are not remotely the same thing.

‍

Questions to Ask Before the Next Growth Push

Before you approve the next acquisition plan, ask:

1. How much of reported online conversion is truly self-serve?
2. Are support teams completing orders inside the same system the customer uses?
3. Have internal and support-assisted actions been filtered cleanly enough to trust the baseline?
4. Can you see where people are dropping out of the journey without guessing?
5. Are friction priorities coming from behavior or from whoever sounds most confident in the meeting?
6. Has the website earned more traffic, or has the business simply gotten good at covering for it?

Those questions decide whether the next dollar belongs in scale or in finally diagnosing what the reporting has been hiding.

‍

Before You Trust the Dashboard Again

What makes the Farmers Depot case worth paying attention to is not that it turns analytics cleanup into some heroic transformation story.

It is useful because it names a failure pattern that shows up in more businesses than people want to admit.

A site can be profitable and still be misread.

A conversion dashboard can be technically accurate and strategically misleading.

And when customer service is carrying a large share of the ordering burden, the website can look healthier than the self-serve experience actually is.

Before you fund more acquisition, get clear on how much of your "online" conversion is actually online in the way you mean it.

If that line is blurrier than it should be, start there. Not with a bigger media budget. With a harder, cleaner read on how much of the funnel belongs to the website and how much of it is still being held together by people behind the scenes.