Your Meta dashboard says one thing. Shopify says another. Your agency report somehow says both are a win.
That gap is where bad decisions get made.
Meta attribution is Meta’s attempt to assign credit for a sale to an ad someone saw or clicked. It is useful, but it is not a clean, final answer to the question every founder actually cares about: did we profitably create more demand, or did we just claim credit for an order that was already coming?
If you are spending serious money on Meta, treating Ads Manager as the source of truth is how you end up scaling a campaign that looks brilliant in-platform and quietly erodes margin in the bank account.
What Meta attribution actually measures
Attribution is a set of rules. The rules decide which marketing touchpoint gets credit when a customer buys.
Meta typically reports conversions against a selected attribution setting, commonly a seven-day click and one-day view window. In plain English, Meta can claim a purchase when someone clicked an ad within the last seven days, or viewed one within the last day, then converted.
That does not mean Meta fabricated the sale. It means Meta had contact with that customer within its reporting window. Those are very different claims.
Consider a shopper who sees your video ad on Tuesday, searches your brand on Google on Friday, reads reviews, then buys on Sunday using an email discount. Meta may report the sale. Google may report it. Your email platform may report it. Shopify may credit the last click to Google or email. None of these systems is necessarily lying. They are using different rules to describe the same customer journey.
The problem starts when a media buyer uses Meta-reported ROAS as the only measure of performance. That number is a directional signal, not permission to spend without restraint.
Why Meta attribution and Shopify rarely match
Founders often assume a discrepancy means tracking is broken. Sometimes it is. More often, it is the expected result of platforms measuring different things.
Shopify generally leans towards last-click attribution in its standard reporting. Meta uses its own pixel and Conversions API signals, applies its attribution window, and can model some conversions when direct tracking is unavailable. Add consent choices, iOS privacy restrictions, cookie expiry, cross-device browsing and customers who return through a bookmarked page, and a perfect match becomes unrealistic.
The goal is not to force the numbers to match. The goal is to understand what each number is telling you, then make commercial decisions from the full picture.
If Meta reports $100,000 in purchase value and Shopify records $70,000 attributed to Meta, do not immediately declare that Meta is overstating by 43 per cent. Check whether total store revenue grew, whether blended customer acquisition cost held, whether new-customer sales increased, and whether branded search or email revenue moved at the same time.
Meta can influence demand that finishes elsewhere. It can also overclaim demand created by your existing brand, email list, organic content or repeat customer base. Both can be true in the same account.
The attribution mistakes that waste ad spend
Most underperforming accounts do not have an attribution problem. They have a decision-making problem disguised as an attribution problem.
The first mistake is optimising to reported ROAS without separating new and returning customers. A brand with a strong repeat base can show a healthy Meta ROAS while ads primarily harvest people who already know the product. That may keep the dashboard pretty, but it does little to expand the customer file.
The second is treating view-through conversions as equal to click-through conversions. A one-day view can be valuable for a low-consideration product with a short buying cycle. For a higher-ticket product or a purchase with a longer decision period, it deserves more scepticism. Look at it, but do not build your budget around it.
The third is changing too many inputs after a soft week. Attribution lag means today’s reported revenue is incomplete. If you slash budgets, replace creative and rebuild audiences every few days, you destroy the conditions needed to learn what is working. Responsiveness is good. Panic optimisation is expensive.
The fourth is using blended ROAS as a hiding place. Blended efficiency matters because payroll and stock are not paid with platform metrics. But a healthy blended result can conceal a weak acquisition engine if branded search, email and returning customers are carrying the load. You need both the business-level view and campaign-level diagnosis.
The numbers worth watching instead
A useful measurement framework has layers. Start with total Shopify revenue, gross margin, contribution margin after advertising, and new customer volume. These tell you whether the business is actually moving forward.
Then assess blended metrics such as marketing efficiency ratio and blended customer acquisition cost. They reduce the temptation to let Meta, Google and email each take credit for the same dollar.
Finally, use Meta’s own reporting to make tactical choices. Which creative earns attention? Which offer converts? Which campaign is finding buyers at an acceptable cost? Platform attribution is particularly useful for comparing ads inside the same account, provided the setup is consistent.
For a founder-led Shopify brand, the question is not, “Which dashboard is right?” Ask, “What decision can this dashboard help us make?”
Meta data can tell you whether Creative A is outperforming Creative B under the same conditions. Shopify can tell you whether revenue is rising. Your customer data can tell you whether growth is coming from first-time buyers or people who were likely to purchase anyway. Together, that is a far more useful picture than arguing over a reporting gap.
Set up Meta attribution before judging performance
Before you decide an account has a traffic or creative problem, confirm the measurement foundation is sound.
Your Meta Pixel and Conversions API should both be active, deduplicated correctly and firing standard ecommerce events with accurate value and currency data. The purchase event needs to trigger once per completed order, not twice because browser and server events are not being matched. Product catalogue data, domain verification and event prioritisation also need to be clean.
Check the basics in Shopify as well. Are test orders excluded? Is your checkout domain configured properly? Are refunds, subscriptions and post-purchase upsells being interpreted consistently? A headline ROAS is worthless if the purchase value feeding it is wrong.
Then choose attribution settings that match how people buy from you. There is no universal setting that makes reporting truthful. A $35 impulse product and a $900 considered purchase do not deserve identical assumptions. Your job is to use a consistent lens long enough to compare performance, while validating the results against real business outcomes.
How to test whether Meta is creating incremental revenue
Incrementality is the uncomfortable question: would these sales have happened without the ads?
You will not answer it perfectly from a dashboard. You can get closer with controlled budget tests. Hold creative, offer and landing page conditions as steady as possible, then increase or reduce spend in a defined period and watch what happens to total revenue, new customers and blended acquisition cost. Do not judge the test on Meta-reported revenue alone.
Geographic holdouts can be useful when volume allows. Reduce delivery in a selected region while maintaining it elsewhere, then compare the movement in sales. Audience exclusions can also reveal whether retargeting is merely taking credit for email and direct traffic conversions.
These tests have trade-offs. Small brands often lack enough volume for statistically tidy experiments, and seasonality can muddy the result. That is not a reason to avoid testing. It is a reason to make fewer, cleaner changes and interpret results with discipline.
A practical approach is to keep a simple weekly scorecard: total revenue, ad spend, blended efficiency, new customer revenue, Meta-attributed purchase value and major changes made in the account. After several weeks, patterns become harder for a polished agency report to hide.
Use attribution to make better calls, not prettier reports
When Meta’s reported ROAS falls, do not automatically cut spend. Check whether total sales and new-customer acquisition remain profitable. You may be reaching colder audiences, which can lower platform efficiency while strengthening the business.
When Meta’s ROAS rises, do not automatically scale. Check whether frequency is climbing, whether retargeting is doing the heavy lifting, and whether branded demand or an email promotion explains the lift. Scaling a misleading signal is one of the fastest ways to turn a good month into a bad quarter.
This is where experienced account management earns its keep. The job is not to recite attribution caveats until nobody is accountable. The job is to build a measurement system that is honest about uncertainty, then make decisive budget, creative and audience calls anyway.
Founders do not need perfect attribution. They need enough clarity to stop funding stories and start funding profitable growth. Keep your eyes on contribution, new customers and total Shopify revenue. Let Meta reporting guide the next move, but never let it write the verdict.