The only ads that sit above the ads.

A charge happens when somebody phones, messages or books through the ad. Browsing costs you nothing. That sounds strictly better than paying per click and it moves the risk rather than removing it, because the thing you now manage is lead quality rather than click quality.

Google started this as Home Service Ads in a San Francisco pilot in 2015 and spent years widening it. What is genuinely recent is availability: the expansion across metros and into categories beyond home services mostly landed in the last two or three years, which is why plenty of businesses only got the option lately and why it still feels like a new product.

License, insurance and a background check on the business and often its owners. For home services the badge is Google Guaranteed; for professional categories such as legal and financial it is Google Screened. The verification is the product, and it is the barrier that makes the placement worth something.

You choose service categories and a service area. No keyword list, no match types, no negative list, no ad copy. Your listing is the business name, the rating, the review count, the hours and a badge, which is a much smaller surface to optimize and a much less forgiving one.

Calls and texts arrive as leads rather than as form fills, so in practice this needs a CRM behind it to track the conversation, hold the history and make sure nothing sits unanswered. We run ours on HighLevel. Without something in that slot the leads are in an app on somebody's phone, which works right up until it does not.

Wrong service, outside your area, spam, a duplicate of one you already paid for: all of these can be disputed and credited. It is the most neglected part of running LSA and the difference between a real cost per lead and a flattering one.

A charged phone lead came out at 1.88 times the cost of a charged message lead in the first account and 1.76 times in the second. Two different accounts, two different periods, near enough the same answer, which is what makes it worth writing down rather than treating as a quirk of one month.

The lead mix ran 51% phone and 49% message in the first account, 54% and 46% in the second. So calls and texts turn up about as often as each other, which is not what most owners expect and is worth knowing before you decide which one to staff for.

Phone leads took 66% of the spend in the first account and 67% in the second, for about half the leads. That single sentence changes how you think about the answering problem: the expensive half of your lead flow is the half that hangs up if nobody picks up.

That text leads are better value. They are cheaper per lead and we have no close rate attached to them here, so anybody telling you which converts better is guessing. What it does justify is measuring the two separately from the day you switch on, which most accounts never start doing.

Proximity does much of the work, so a service area drawn too wide quietly under-performs everywhere in it. Tightening the area usually raises the rank inside the part you actually wanted.

Ten reviews is the gate: you cannot open a Local Services Ads account without them, so for a business sitting on six this is the first job rather than a later one. Past the gate they keep working, because the score and the count both feed the ranking and the reviews on your Business Profile are the ones that count. Review generation stops being a nice-to-have and becomes part of the media buy.

Response rate and speed to calls and messages feed the ranking directly. A business that misses calls at four in the afternoon is being demoted for it and usually has no idea that is what is happening. This is also why the CRM matters more here than the ad settings do.

The panel favors businesses open when the search happens. Listing hours you do not really cover buys you leads you then answer badly, which costs twice.

Serious or repeated complaints affect standing and can cost the badge outright. The verification that makes this placement valuable is the same verification that can be withdrawn.

Budget governs how much of the available demand you can take rather than where you sit against a rival with better reviews and a faster phone. Raising it does not buy you past the proximity and reputation signals, which is a hard conversation to have and an honest one.

Google lets you mark each lead as it comes in — booked, not a fit, spam. That is a per-lead click by a human being, and it is the input the dispute process and your own reporting both rest on. At a low lead volume it is fine. At a few hundred a month it stops happening, and once it stops the reported cost per lead drifts away from the real one.

Leads you should not have paid for are only credited if somebody raises them. An account nobody grades is an account quietly paying for wrong-area calls and spam texts every week, and the number never appears anywhere as a loss.

Grading and tracking driven off the CRM record rather than off somebody remembering to click, so the lead is scored by what actually happened on the call or in the thread. We have it running on HighLevel and are extending it. It is not a Google feature and we do not expect one.

Grade for one full week and count what you would have disputed, then multiply. Most owners we do this with are surprised by the number, and it is usually enough to justify putting a process behind it.