• Shopware Gold Partner & Shopify Partner

Chapter 2

Four myths that cost managing directors money

Table of contents

Four widespread assumptions lead managing directors to steer their ad budget by a false picture. So before we get to the twelve metrics you really need, there is a more fundamental question: What do the numbers in your dashboards actually show?

These four myths are behind many of the most expensive wrong decisions in e-commerce marketing. If you do not see through them, you optimize for the wrong numbers, however good your methodology is afterwards.

Myth 1: "The conversion value in Google Ads is my ad revenue"

The conversion value in your Google Ads account is Google Ads' attribution, not the additional revenue in your store backend. Google Ads claims a purchase if the buyer interacted with an ad within the conversion window (often 30 days, configurable up to 90 days). This applies even if:

  • the buyer already knew your brand and simply clicked on the ad after a brand search,

  • the real trigger came from the newsletter, organic search or a recommendation – Google Ads does not see these touchpoints at all,

  • the same purchase is also counted as a success by Meta or an affiliate network.

That is not a mistake on Google's part but a matter of perspective: every platform measures its share of the purchase paths it was involved in. What it does not measure is incrementality, i.e. the revenue that would not have happened without the ad. How large the gap between the two is varies from store to store. With a high share of brand searches and returning customers, it is typically large.

What data-driven attribution has to do with it

Since 2023, Google Ads has offered only two attribution models: last click and data-driven attribution (DDA). First click, linear, time decay and position-based were retired, and DDA is the default. DDA distributes a conversion across several Google ads on a purchase path, for example 0.4 to a Shopping ad and 0.6 to a search ad. In total, it remains one conversion.

So DDA does not inflate the total, it redistributes it. With native Google Ads tracking, this happens within the Google ad touchpoints. If conversions are imported from GA4, other channels such as organic search can also receive a share, depending on the settings.

The gap to additional revenue arises elsewhere: from touchpoints Google Ads does not see, from long windows and above all from purchases that would have happened without the ad anyway. Whether a conversion is observed or modeled (Myth 2) changes nothing about this: an attributed conversion does not prove an additional purchase. DDA helps automated bidding weight the ads on a purchase path better. It does not answer the question of additional revenue.

What follows from this

A typical sequence: the report shows €380,000 in conversion value on €34,000 of ad spend, ROAS 11.2. On that basis, the budget is doubled. Three months later, store revenue is up 15 percent. Sounds like success. The math behind it (simplified, with a 40 percent contribution margin before advertising):

Before

After

Change

Store revenue per month

€540,000

€621,000

+€81,000

Contribution before advertising (40%)

€216,000

€248,400

+€32,400

Ad spend

€34,000

€68,000

+€34,000

Contribution after advertising

€182,000

€180,400

–€1,600

15 percent more revenue, and in the end less money is left over. This comparison does not show whether the advertising worked or whether diminishing returns kicked in. But it does show that the reported conversion value is no proof of additional, profitable revenue.

The only relevant question when someone presents you with a conversion value is: How much of it arrived in the store backend as additional revenue that would not have happened without this campaign? That can only be answered with a test, such as a controlled pause of brand campaigns, a regional comparison test or a conversion lift study.

Myth 2: "The numbers in my dashboard are measured"

Some of the conversions in your Google Ads account were not observed but estimated. Google fills in conversions that cannot be directly attributed to an ad using machine learning and reports both in the same column. This mainly concerns three cases:

  1. Declined consent: If a visitor declines the cookie banner, Google uses Consent Mode to estimate how many conversions these visitors generated.

  2. Device switching: Click on the smartphone, purchase on the laptop. Google models the connection.

  3. Browser restrictions: Where Safari or ad blockers prevent attribution, Google also models.

According to Google, modeled conversions are only included when the system is confident about their quality. For you as managing director, one consequence remains: the standard report shows a single number. You cannot see there which share is observed and which is estimated. The share depends mainly on your consent rate and can change over time without you noticing. A drop after a banner or browser update is therefore not always a drop in the business.

Estimated does not mean worthless. But for your budget decision, you need to know whether the business has changed or only the measurement.

Myth 3: "I can add up the ad revenue of my channels"

The same purchase is counted as a success by several platforms at once. A typical purchase path: on Monday the customer clicks a Meta ad, on Wednesday they click a Google ad after a brand search, on Friday they come to the store directly and buy.

  • Meta counts the purchase because the click falls within its own window (by default seven days after a click).

  • Google Ads counts the purchase because the click falls within its own window.

  • Your store backend records the purchase once, with the real amount.

Each platform correctly measures what it is supposed to measure. The problem only arises when someone adds up the numbers:

Channel

Ad revenue according to the platform

Google Ads

€380,000

Meta

€220,000

Affiliate

€90,000

Total

€690,000

Actual store revenue in the same month: €540,000. The sum of ad revenue is 28 percent higher than total store revenue, even though some of the orders did not come from advertising at all. The longer the purchase path and the more channels involved, the greater the overlap.

There is only one number without overlap: revenue in the store backend. Every order is recorded there exactly once. This gives you a simple, robust metric for overall steering: the MER (marketing efficiency ratio) = store revenue ÷ total ad spend. It shows your overall efficiency without double attribution across channels. Whether that efficiency is profitable is decided – as with ROAS – by your margin: if your MER is above your break-even ROAS from Chapter 4, your ad budget as a whole is making money.

Myth 4: "With server-side tracking, I see 100 percent again"

For most mid-sized stores, server-side tracking is a worthwhile investment. It makes data transmission more robust, reduces losses from browser restrictions and ad blockers and gives you back control over your data. Four gaps remain structurally, however:

Gap

Why server-side tracking does not close it

Declined consent

Anyone who declines may not be tracked for advertising purposes – whichever route is used.

Modeling at Google

Once data arrives at Google, the same estimation methods run as before.

Double counting across platforms

Each platform still claims its own share.

Bots and false measurements

Without clean filtering, a server setup can even contain more false measurements.

Server-side tracking significantly improves data quality. It does not replace the method you use to interpret the numbers. Anyone who blindly trusts their dashboards after the switch repeats the original mistake at a higher technical level.

What the switch looks like in Shopware and Shopify, what it delivers and what it does not, is covered in the guide Server-Side Tracking for Shopware and Shopify.

What you can check today

  1. Ask your marketing team: how much of the conversion value in Google Ads is additional revenue that would not have happened without advertising – and how was that measured?

  2. Check which attribution model and which conversion window are set in your Google Ads account.

  3. Put the sum of ad revenue across all platforms next to store revenue for the same month.

  4. Calculate your MER – store revenue divided by total ad spend – for the last twelve months and compare it with your break-even ROAS.

  5. If you use or are planning server-side tracking: how exactly will the success of the switch be measured?

Frequently asked questions

No, but it answers a different question. It shows which purchases Google ads were involved in. It does not show how much revenue would not have happened without ads.

A model that distributes one conversion proportionally across several Google ads on a purchase path. The total remains one conversion. With native Google Ads tracking, only Google ad touchpoints are involved. For conversions imported from GA4, other channels can also receive a share, depending on the settings.

Conversions estimated by Google that cannot be observed directly. For example after declined consent or a device switch. They appear in the same column as observed conversions.

Only with tests. Proven methods are controlled pauses of individual campaigns, regional comparison tests and conversion lift studies. Without a test, every figure for additional revenue remains an assumption.

The marketing efficiency ratio: store revenue divided by total ad spend. It needs no attribution and shows the overall efficiency of your ad budget. Whether it is profitable is decided by your margin: if your MER is above your break-even ROAS, your ad budget as a whole is making money.

For most stores with a significant ad budget, yes. It noticeably improves data quality and therefore automated bidding. It just does not bring back 100 percent and does not replace correct interpretation.