HowtoFeedMarginDataIntoGoogleAdsConversionValues

The single highest-impact technical move for any ecommerce Google Ads account is feeding profit data into Smart Bidding as the conversion value, rather than letting Google optimise against revenue by default. This article shows the three ways to do it (conversion value rules, offline conversion imports, server-side custom values at conversion time), the trade-offs of each, a step-by-step setup for the most reliable approach for most £10k+/month accounts, and how this feeds into the BOI™ framework for SKU-job bidding.

For context on why profit signals beat revenue signals, see our POAS framework explainer and our POAS vs MER vs ROAS guide. For the upstream data work (calculating true profit per order in Shopify), start with POAS calculation for Shopify brands.

Why default revenue-based conversion values lead to bad bid decisions

Google Ads Smart Bidding optimises against whatever signal you give it. By default, the conversion value passed into Smart Bidding is revenue — the order total Shopify reports back when a conversion fires.

The problem: revenue is not profit. A £100 order on a low-margin product generates the same conversion value as a £100 order on a high-margin product, even though one delivers £15 of contribution margin and the other delivers £70. Smart Bidding can't tell the difference. It pursues whichever campaign drives more revenue per click, regardless of whether the underlying margin justifies the bid.

For brands with uniform margins across the catalogue, this barely matters — every £100 of revenue is roughly £25 of margin, the relative bid logic works out. For brands with margin variation (which is most ecommerce brands), it matters enormously. The bid signal systematically over-rewards low-margin bestsellers and under-rewards high-margin niche lines. Over months, the campaign mix shifts toward unprofitable products.

The fix: pass profit (or margin-adjusted revenue) into Google Ads as the conversion value instead of raw revenue. Three implementation routes. Each with trade-offs.

Option 1: Conversion value rules in Google Ads (fastest, blunt instrument)

What it is: Google Ads' built-in conversion value rules let you apply a multiplier to the reported conversion value based on conditions you define. You can set rules by location, device, audience, or — most usefully for margin work — by category or custom label.

How it works for margin: Tag your products with custom labels in your Merchant Center feed by margin band (e.g., "high-margin," "medium-margin," "low-margin"). In Google Ads, set conversion value rules that multiply revenue by a margin factor for each band — e.g., low-margin × 0.25, medium-margin × 0.45, high-margin × 0.65. Smart Bidding now sees the margin-adjusted value rather than raw revenue.

Setup time: 2–4 hours assuming custom labels are already populated.

Trade-offs

Pros: Fast to set up. Doesn't require any data pipeline work. Operates in real-time at bid time. Reversible if you mess it up.

Cons: Only as granular as your custom label bands (typically 3–5 bands across the catalogue). Misses SKU-level margin variation within bands. Doesn't account for variable costs that vary per order (shipping, returns).

Best for: Brands with broadly clustered margins (e.g., apparel with mostly 50–60% margins, a few outliers) where 3–5 margin bands capture most of the variation. £10k–£50k/month spenders looking for a quick win.

Option 2: Offline conversion imports (most common, weekly cadence)

What it is: You calculate true contribution margin per order after the fact (weekly or monthly), then upload the corrected conversion value back to Google Ads via Offline Conversion Imports. Google Ads replaces the original revenue-based conversion value with your true margin figure.

How it works for margin: Export your orders weekly from Shopify, join to cost-of-goods at the variant level, subtract shipping/fees/discounts, calculate true contribution margin per order. Upload to Google Ads as offline conversions matched on Google Click ID (GCLID). Smart Bidding's historical learning gets re-weighted toward margin, and going-forward bids adjust accordingly.

Setup time: 1–2 weeks for initial setup (GCLID capture in Shopify, data pipeline to join Shopify and Google Ads data, upload process). Ongoing: 1–2 hours per week to run the import.

Trade-offs

Pros: Per-order accuracy. Captures variable costs (shipping, returns) that conversion value rules can't. Established methodology — well-documented across the industry.

Cons: Requires GCLID to be captured on every order (Shopify allows this with the right setup). Requires a data pipeline to join Shopify and Google Ads data. Bidding signal is lagged — the algorithm learns from corrected values, but with a week or two of delay.

Best for: Brands spending £20k+/month who want full accuracy and have the operational capacity to run weekly imports. The most common choice for mid-size DTC ecommerce.

Setup checklist

1. Enable GCLID capture in Shopify (third-party app or custom code — many free Shopify apps handle this).
2. Build the data join: Shopify order → GCLID → variant COGS → variable costs → true contribution margin.
3. Set up the Google Ads Offline Conversion Import process — either via the manual upload UI or via the Google Ads API.
4. Test with a small batch (50–100 orders) to validate the matching works.
5. Move to weekly cadence once validated.

Option 3: Server-side custom values at conversion time (most precise, requires dev)

What it is: Calculate the corrected conversion value (true contribution margin) at the moment the order is placed, and pass it directly to Google Ads as the conversion value via server-side conversion tracking. No lag, no offline imports, no post-hoc correction.

How it works for margin: When an order is placed in Shopify, a server-side script (typically running in your tag manager or a custom backend) calculates true contribution margin in real-time using variant cost data and a shipping/fees lookup. The corrected value is sent to Google Ads' server-side conversion tracking endpoint. Smart Bidding receives the margin-adjusted value at the same time it receives the conversion fire.

Setup time: 2–4 weeks for initial setup. Requires meaningful development resource. Often involves Google Tag Manager Server-Side or a custom server-side endpoint.

Trade-offs

Pros: Real-time accuracy. No lag. No weekly upload process to maintain. Scales infinitely once built.

Cons: Highest setup complexity. Requires development capacity. Harder to debug when something goes wrong. More moving parts.

Best for: Brands spending £50k+/month with technical resource and a long-term commitment to profit-first bidding. Often best implemented alongside a wider server-side tracking migration (Consent Mode v2, iOS 14 mitigation).

Which option is right for your business

A decision framework:

Most JudeLuxe clients run Option 2. It hits the right balance of accuracy and operational sustainability for the £20k–£100k/month band where most growing DTC ecommerce sits.

Common implementation mistakes

1. Setting up conversion value rules without populating Merchant Center custom labels

The rules need a signal to fire against. If your custom labels aren't populated by margin band, the rules will apply unevenly and produce noisier bid behaviour than the default.

2. Forgetting to handle refunded orders in offline imports

If you upload an order's true contribution margin but never reverse it when the customer refunds, your conversion values overstate profit by the refund value. Build refund handling into the weekly import process from day one.

3. Switching too aggressively from revenue to profit signals

Smart Bidding's machine learning needs time to recalibrate. Switching the conversion value type causes the algorithm to re-learn — bid stability suffers for 2–4 weeks. Phase the change rather than flipping a switch. Run revenue-based and margin-adjusted conversions in parallel for a month, validate the margin-adjusted version matches reality, then swap over.

4. Treating return processing cost as zero

Returns cost money to process even when the original payment fee is refunded. Build a typical per-return cost (£4–£8) into your contribution margin calculation. For high-return categories, this can be 10%+ of the true margin gap.

5. Not testing with a small batch before going live

Always run the import or rule on a small subset first. Validate that the corrected values match what you expect manually. The most expensive POAS optimisation mistakes happen when bad data flows into the system at scale before anyone notices.

How this feeds into BOI™

The BOI™ framework — Bid On Intent — assigns each SKU one commercial job at a time (Profit, Volume, Liquidation, Acquisition) and bids against it dynamically. Feeding margin data into Google Ads is the technical foundation that makes BOI™ work. Without margin-aware bidding signals, every SKU is bid against the same blended signal regardless of its assigned job.

With margin-aware bidding plus BOI™ job assignments, the bid logic becomes:

• Profit SKUs bid against profit-maximisation conversion values (highest margin weighting)
• Volume SKUs bid against volume-favouring conversion values (margin weighted lower, revenue weighted higher temporarily)
• Liquidation SKUs bid against any-conversion values (margin discounted further to maximise throughput)
• Acquisition SKUs bid against predicted-LTV values (first-order margin weighted lower, lifetime margin weighted higher)

The technical step (margin data flowing into Google Ads) and the strategic step (assigning SKU jobs) compound. Either one alone delivers some improvement. Both together deliver materially more.

See BOI™ explained for the full framework.

What to do this week

If you're spending £10k+/month on Google Ads and haven't done this work:

1. Decide which option fits. Use the decision matrix above. For most readers, the answer will be Option 2 (offline imports).
2. Confirm your COGS data is clean. Audit your Shopify variant cost field. Most stores have 5–15% of variants missing or wrong cost data. Fix this first — every other step depends on it.
3. Talk to your dev team or agency about GCLID capture. Most Shopify themes don't capture GCLID by default. Plenty of free apps do it. Get this live before anything else.
4. Build the offline import process — manual upload via UI to start, automate later if needed.
5. Phase the change. Run for 2–4 weeks in parallel with default revenue tracking, validate, then swap over.

By 30 days post-implementation, bid behaviour should be visibly shifting toward higher-margin SKUs. By 60 days, the new equilibrium is established. By 90 days, the POAS picture is reliable enough to make ongoing strategy decisions against.

Frequently asked questions