The most expensive mistake in a B2B SaaS Google Ads account is often not a bad keyword or a wrong bid strategy — it is a setting left on its default. Google Ads ships with a 30-day click-through conversion window, and most accounts never change it. For a business whose buyers take three, six, or nine months to decide, that default quietly throws away the connection between the click that started the journey and the deal that ended it. One industry analysis put it bluntly: this time-horizon mismatch is responsible for more wasted B2B SaaS ad spend than any targeting or bidding error.
This guide covers how the attribution window length actually works, why 30 days starves Smart Bidding on a long cycle, how to choose 60 or 90 days from your own data, and the hard 90-day GCLID constraint you have to design around. It pairs with our deeper look at the conversion window and the SaaS sales cycle; here the focus is specifically on picking a window length and living with the reporting consequences.
Attribution window vs. conversion window
Two settings get tangled together in casual conversation, so it is worth separating them. The conversion window is how long after a click Google will still credit a resulting conversion; for click-through conversions you can set it up to 90 days. The attribution window, as B2B marketers usually mean it, is the lookback horizon over which credit gets distributed across the clicks in a path. For a long-cycle SaaS business they collapse into one practical question: how far back is Google allowed to look when it connects a conversion to an ad interaction?
The reason the distinction matters less than the length is that both fail in the same way when set too short. If a prospect clicks your ad in January, signs up for a trial in February, and converts to paid in April, a 30-day window has long since stopped watching that click by the time the money arrives. The conversion is real, but as far as Google Ads is concerned the January click produced nothing. Multiply that across an account and the platform's entire picture of what works is distorted toward whatever happens to convert quickly — rarely your best, highest-ACV pipeline.
How 30 days starves Smart Bidding
Smart Bidding is only as good as the conversions it can see. The algorithm makes bid decisions based on the conversion data available to it, and it optimizes toward the conversions that fall inside your window. When the window is 30 days and your deals close in month four, the bidder observes that certain keywords generate clicks but few near-term conversions, and it does the rational thing: it bids those keywords down. The keywords it starves are frequently the exact high-intent terms that seed your longest, largest deals — the ones whose value only becomes visible months later.
This is the same distortion we describe in the attribution gap that misleads Smart Bidding, viewed through the lens of time rather than channel. The fix is to widen the window so the algorithm sees the delayed conversions and can reallocate toward the clicks that actually produce them. But widening alone is not enough if the total volume stays thin: Smart Bidding needs a stable base of conversions — a commonly cited floor is 30 or more per month for Target CPA — to bid confidently. A longer window helps most precisely when it lifts previously invisible conversions above that threshold.
Choosing 60 vs. 90 days from your own data
The right window length is a measurement, not an opinion. Pull your own time-to-conversion distribution for whatever event you optimize toward — trial-to-paid, MQL-to-SQL, or booked demo to opportunity — and look at where the bulk of conversions actually land. If most of them close within 60 days, a 60-day window captures them while keeping your reporting reasonably timely. If a meaningful share stretches into month three, extend to the 90-day maximum so you are not clipping the tail that contains your biggest deals.
There is a cost to going long, and it is honesty about timeliness. A 90-day window means recent weeks always look under-reported until late conversions backfill, which can spook anyone reading the dashboard without context. The answer is not to keep the window artificially short; it is to set the window to reality and educate the reporting around it. Choose the length that covers the majority of your real conversion lag, accept that the last few weeks of any report are provisional, and stop optimizing to a 30-day picture of a 90-day business. If your sales cycle genuinely runs past 90 days, the window cannot stretch further — which is where the GCLID constraint below forces a change in what you import.
The 90-day GCLID ceiling and how to work around it
Google Ads caps the click-through conversion window at 90 days for a concrete reason: the GCLID, the click identifier written into your landing-page URL and stored against the lead, expires roughly 90 days after the click. Offline conversion imports depend on that GCLID to reconnect a CRM outcome to the originating click. For a SaaS team whose average cycle is 90 days or longer, this creates a trap: if you wait for closed-won before uploading the conversion, the GCLID may already be dead, and the upload has nothing to attach to.
The workaround is to stop optimizing to the final sale and start optimizing to a strong, earlier proxy that still fires inside the 90-day window. Import the MQL-to-SQL transition, the qualified opportunity, or the booked demo — events that happen while the GCLID is still valid — and assign each a value that reflects its expected contribution to revenue. This is the core idea behind optimizing for SQLs instead of raw leads, and it is why a robust offline conversion stack matters more than window length alone. The window sets how long Google watches; the offline import decides what it gets to see inside that time.
Managing the transition without spooking the account
Changing the window length is not a costless flip of a switch, because it re-bases your numbers and resets what Smart Bidding is learning from. Expect a period where reported conversions shift as the longer lookback fills in, and where recent periods look lighter than the settled ones. Annotate the change in your reporting so anyone reviewing performance understands why the trend line moved, and hold off on drawing conclusions from the first few weeks after the change while the data re-baselines.
Give the bidding algorithm room to relearn, too. When the window widens, Smart Bidding is suddenly working from a fuller and different conversion signal, so a short readjustment period is normal — the same patience you would extend after any material change to the conversion setup. Fold the window review into your regular account hygiene rather than treating it as a one-time fix: as your product moves up- or down-market, your sales cycle changes, and a window that fit a 45-day cycle last year may clip a 75-day cycle this year. Getting this setting right is unglamorous, but for a long-cycle B2B SaaS account it is frequently the highest-return change available, because it fixes the data every other optimization depends on.