Google Ads now offers two attribution models: data-driven (the default) and last click. Data-driven attribution spreads conversion credit across the Google ads a buyer clicked before converting, weighted by what your own account data says made a difference. Last click gives 100% of the credit to the final ad clicked. For most B2B SaaS accounts, data-driven is the right model for the conversion actions Smart Bidding optimises toward. Neither model sees your sales cycle or other channels, so pipeline reporting has to come from the CRM.
What an attribution model does in Google Ads
An attribution model decides which ad, keyword and campaign get credit when a conversion happens after more than one ad interaction. That credit has two jobs: it fills the Conversions column you read in reports, and it is the signal Smart Bidding uses to decide what each keyword is worth in the next auction. Change the model and you change where the algorithm spends.
One limit matters more than the choice of model: Google Ads only shares credit between Google ad interactions (Search, YouTube, Display, Shopping, Demand Gen). A buyer who heard about you on a podcast, clicked a LinkedIn ad and then searched your brand on Google is a 100% Google Ads conversion as far as Google Ads is concerned.
Data-driven vs last click: how they differ
| Data-driven attribution | Last click | |
|---|---|---|
| How credit is assigned | Fractional, across the Google ad clicks on the path, based on your account's converting vs non-converting paths | All credit to the final Google ad clicked |
| Default for new conversion actions | Yes | No |
| What it tends to over-credit | Can spread credit to clicks that were incidental when paths are short and data is thin | Brand search and remarketing, the usual last touches |
| What it tends to under-credit | Little, within Google; nothing outside Google is seen at all | Category, competitor and YouTube campaigns that start the journey |
| Best use in B2B SaaS | Primary conversions that Smart Bidding optimises to | Simple reconciliation with a last-touch CRM report |
First click, linear, time decay and position-based attribution have been removed from Google Ads, and conversion actions that used them were moved to data-driven. The timeline and what to check after the migration are in Google Ads removes four attribution models.
Why the choice matters more for SaaS than for ecommerce
A typical B2B SaaS path looks like this: a category search (“customer onboarding software”), a comparison search a week later, then a brand search the day the buyer books a demo. Under last click, the brand campaign gets the demo, and the category and comparison campaigns that created the brand search get nothing.
That has two consequences a growth lead feels directly:
- Generic campaigns look unprofitable and get cut. Brand volume then falls a month or two later, and nobody connects the two because brand still shows a great cost per demo.
- Smart Bidding under-bids the top of the journey. The algorithm learns that category keywords rarely convert, bids them down, and the account shrinks into brand and remarketing.
Data-driven attribution doesn't fix measurement outside Google, but within Google it gives category and competitor clicks a share of the demo, which keeps the bidding honest about what starts the journey.
Which model a B2B SaaS should use
| Situation | Model | Why |
|---|---|---|
| Primary conversion actions used for bidding (demo, SQL, opportunity imports) | Data-driven | Gives the campaigns that start the journey credit, so Smart Bidding keeps funding them |
| Secondary / observation-only actions | Either | They don't steer bidding; pick whichever matches how you report |
| Account with very few multi-click paths (mostly brand) | Either, results will look similar | With one click per path, the models assign the same credit |
| Board or finance reporting | Neither | Use CRM pipeline with the click ID on each record, plus incrementality tests |
To check or change the model: Goals → Conversions → Summary, open the conversion action, edit settings, and set Attribution. Before you switch a primary action, open the Model comparison report under Attribution to see how much credit would move between campaigns, and expect a short Smart Bidding adjustment period afterwards.
The gap neither model closes: your sales cycle
Both models only attribute conversions that Google Ads records. For a sales-led SaaS, the conversion that matters — a qualified opportunity or a closed deal — happens in the CRM weeks or months after the click. Two settings decide whether Google ever sees it:
- Offline conversion imports. Without SQL, opportunity and closed-won stages flowing back with the click ID, both models are attributing demo requests, not revenue. See how to measure pipeline from Google Ads.
- The conversion window. It can be set from 1 to 90 days for click-through conversions. A stage imported after the window has closed is not attributed at all. Match it to your click-to-SQL time, as covered in why the 30-day window breaks B2B SaaS attribution.
Platform attribution, CRM attribution and incrementality
Funded SaaS teams usually end up arguing about attribution because one number is being asked to do three jobs. Separate them:
- Platform attribution (data-driven in Google Ads) steers bidding. Judge it on whether it makes the bidding smarter, not on whether its totals match the CRM.
- CRM attribution reports pipeline and revenue by source to the board, using the click ID stored on each lead.
- Incrementality answers “what did the spend cause?” when you are deciding budgets. Use conversion lift or geo holdouts, and marketing mix modeling once spend is large enough.
Check your own setup
If your bid strategy is optimising to demo requests on last click with a 30-day window, it is probably under-funding the campaigns that create your pipeline. A B2B SaaS Google Ads audit checks attribution, conversion windows and offline imports together, or start with the free 10-point Google Ads audit.