The question arrives around day eleven, usually as a text message sent at nine at night. How long does it take for Google Ads to work is the polite phrasing. The version we get is closer to “is this thing on.” The answer that has held up across our own onboardings is that the first three weeks produce the worst cost per booked job you will see all year, weeks four through eight are where the money comes back, and week eight is the first checkpoint worth arguing over. Anything before that is weather.
Three weeks, or one to two sales cycles, whichever runs longer
Google publishes the number and it is worth reading in the original. After a change, its documentation says, “It can take up to 3 weeks or 1-2 conversion cycles for the bid strategy to calibrate to the new objective, although it can be faster depending on the amount of conversion data present” 1.
The phrase doing the work there is “conversion cycles,” and it is not a calendar. A conversion cycle is the time between someone clicking your ad and doing the thing you count as a win. If you sell same-day drain clearing, that gap is a few hours and three weeks is a realistic clock. If you sell a $40,000 commercial install that takes a homeowner two months of quotes and a spouse conversation, then one to two conversion cycles is two to four months, and no amount of budget compresses it. What the system is waiting on is your customers making up their minds.
The same page adds a line that rarely makes it into a client call: “Our algorithms continue to learn even when the bidding status no longer shows ‘Learning’” 1. So the status label clearing is not a finish line. It is the platform saying it has enough data to stop flailing.
Weeks one to three: the number is supposed to look bad
Expect your cost per booked job in this window to be the highest figure the account will produce, and expect it to be well above whatever range you were quoted. A home services account we take over commonly starts around $260 to win one booked job. That is the price of the platform buying information about which searches turn into work, and you are paying for the education whether or not anyone told you in advance.
The mistake in this window is touching things. Practitioner guidance from the same publisher that produces the widely used search benchmarks holds that most campaign changes take about 30 days to settle 2, which means an owner who adjusts bids on Monday and budgets on Thursday has restarted a clock they were already impatient about. Daily fiddling produces an account permanently in week one.
Do one thing instead during those three weeks: log every inquiry and mark whether it turned into work. You will need that list, and nobody else can build it for you.
Weeks four to eight: the wasted spend comes out
This is where the largest single improvement lands, and the mechanism is dull. By week four there is enough history to see the actual searches that triggered your ads, and a meaningful share of them are wrong. Job seekers. Students writing papers. People shopping for a part rather than a service. Neighboring towns your crew will not drive to. Cutting those is where most of the gain in the first quarter comes from, and it is a reading exercise before it is a technical one.
Across our own onboardings the shape has been consistent enough to plan around: a home services account that started near $260 per booked job settles somewhere between $150 and $180 by the end of month three. Nobody found a clever setting to get there. A third of the spend was going somewhere it should not have been, and it stopped going there. You can see the same pattern with the account details changed in our case studies.
Weeks nine to twelve: the page and the offer
The account stops being the constraint around week nine, and whatever happens after a click becomes the whole game. Changes to the landing page, the form, the phone script, and the offer show up in this window, and their effect is larger than anything left to do inside the ad platform.
That ordering matters for how you judge a provider. If somebody spent weeks one through eight rewriting your page and never opened the search terms report, they worked in the wrong order and your money paid for the detour.
What does not improve on schedule
Three things sit outside the ramp entirely, and pretending otherwise is where a lot of disappointment comes from.
Your close rate does not improve because the ads got better. If your team books 15% of the calls that come in and a competitor books 35%, you are paying roughly twice what they pay for the same booked job at identical ad performance, and that gap survives any amount of account work.
Seasonality does not care about your start date. Launch an HVAC account in October and month three lands in December, so the ramp and the season are moving in opposite directions and you cannot tell them apart without last year’s numbers.
And the auction itself can simply be unaffordable in your market at your price point. When that is the finding, the correct outcome is to stop, which is why our ad spend calculator will tell you no and show a six-month ramp rather than a hockey stick.
What owners ask around week six
Can I speed this up by raising the budget? Somewhat, if the account is starved. The clock in Google’s documentation runs on conversion cycles, so more budget means more data per week and a faster calibration, up to the point where you are buying searches with weaker intent. Doubling a $1,500 budget in week two usually buys you a larger sample of the traffic you were about to cut anyway.
We changed our budget in week three. Did we restart the clock? Probably part of it. Any change to the objective or the bidding puts the strategy back into calibration, and with roughly 30 days for changes to settle, three edits in a month means an account that has never had a clean read 2.
Week eight looks bad and you are my agency. What should I ask you? Ask for the search terms report with the negatives added and the dates they were added. Then ask which week we changed the landing page and what the booking rate did after. If we cannot name the specific changes we made in weeks four through six and what each one moved, you are paying a retainer for a dashboard, and you should say so on the call. Our fees are published on the services page, which makes that conversation easier to have.
The week-eight test
An agency promising results in the first 30 days is promising to move a number that can be improved without helping you, because clicks and cost per click both look better the moment someone bids on cheaper traffic, and cheaper traffic is how you end up with a busy dashboard and a quiet phone. Ask what the 30-day promise is measured in. If the answer stops at clicks, impressions, or form fills, what you have been sold is a report on the platform’s behavior with your name on the cover.
Here is the test to run yourself, and it takes ten minutes. Add up total ad spend for weeks one through three and divide by the number of jobs you booked out of it. Do the same for weeks five through eight. If the second number is not lower than the first, the learning period is no longer the explanation. Our read is that the cause at that point is a page or a phone rather than a bid: a form asking for eleven fields, a call ringing out at 4:40 on a Friday, an offer priced above the market you are advertising into.
Run both numbers before your next invoice is due, and open the call with those two figures.