A conversion count is not a fact about the market. It is a clock you chose, and Google just gave you every integer day under that clock.
Your conversion count is a setting.
On August 11, Google Analytics made click-through conversion windows any integer from 1 to 90 days, and engaged-view windows any integer from 1 to 30 days (Google Analytics Help, PPC Land). Engaged-view had been locked at three days for every property. Click-through had been a six-step ladder. Those constraints are gone.
That sounds like a UI nicety. It is not. Conversion volume feeds automated bidding. Change the window and you change the score the optimizer chases, and the board slide that claims "conversions are up."
Write The Cycle Before You Touch The Control
Pick the conversion that actually runs the business: demo booked, qualified form, checkout. Write how many days a real buyer usually takes from the paid click or the video view to that event. Use last-quarter CRM, not a round number that feels tidy.
Then open Advertising > Conversion management > Settings (or the linked Google Ads conversion screen) and set that integer. Do not leave engaged-view at three days because "that is what it always was." Do not leave click-through at 30 because the old dropdown stopped there.
Freeze It Before You Plan Q4
A window you keep editing is a moving scoreboard. Set it once for the planning cycle. Put the date and the day-count in the measurement notes the CMO and the media lead both see.
Assume the change applies forward only until Google says otherwise. Year-on-year comparisons that cross the edit date compare two definitions. Annotate the chart or do not use that comparison.
What This Is Not
This is not a reason to lengthen every window to 90 so the dashboard looks healthier. A longer window reassigns credit. It does not invent demand. If video suddenly "drives" more conversions after you move engaged-view from 3 to 21, ask whether search and direct lost assists they were already getting.
The operating move is boring on purpose: one cycle written in days, one window set to match, one freeze date, one annotated chart. Until that is true, every efficiency claim is an argument about a setting, not about the market.
