28% of the budget, spent with the door closed.
PPC audit for Sablare Laser, a workshop for laser cleaning, mechanical blasting, painting and refurbishment in Cernica, Ilfov. The campaigns were run by another agency; the website had just been rebuilt.
- Client
sablarelaser.ro - What we audited
- Google Ads (12 months of the account) · conversion tracking · landing pages against the new site (15 service pages)
- Campaigns run by
- Another agency. The audit delivered the work list.
- Service
- PPC audit →
What did the PPC audit find at Sablare Laser?
That 69% of paid clicks landed on the homepage even though the new site had 15 service pages; that 28% of the budget was spent in the evening and at weekends, when nobody answered the phone; and that 88% of “conversions” were simple taps on the phone number.
Sablare Laser had a new website with a dedicated page for each service, and an ad account built for the old one.
The audit read 12 months of the account — landing pages, budget by service, delivery schedule, what counts as a conversion, audiences and negative keywords — and was presented in a joint meeting with the client and the agency that runs the campaigns. The work list was accepted in that same meeting.
of paid clicks landed on the homepage, even though the new site had 15 service pages.
of the budget was spent in the evening and at weekends, when nobody answered the phone.
of “conversions” were simple taps on the phone number, counted more than once.
How we run a PPC audit
- 01
The account, piece by piece
Landing pages, budget by ad group, hours and days, conversion types, audiences, negative keywords — each with its own figure.
- 02
Read-only access
We changed nothing in the account.
- 03
The account against the new site
Which pages exist now on the new site (built by Websem), which services are no longer offered, which addresses disappear at launch.
- 04
A joint presentation
The client and the agency running the campaigns, in the same meeting, with decisions taken on the spot; then a written brief for the agency: 5 campaign structures, ~150 negative keywords by category, tracking instructions.
Six findings, from the landing page to the negative keywords
01 · The pages
The person searched for one specific service. They landed on a page about everything.
Three destination addresses were used across the whole account. The new site had a dedicated page for each service — and at launch, the old addresses were going to disappear.
What worked
The two main services already had their own page and a quality score of 7–8.
What we recommended
One ad → the page of the service searched for; redirects and new addresses in the ads on the same evening as the launch.
- Homepage69%
- Blasting page20%
- Painting page11%
Chart data
| Landing page | % of clicks |
|---|---|
| Homepage | 69% |
| Blasting page | 20% |
| Painting page | 11% |
02 · Budget by service
of the budget went to a service that was no longer offered.
A single budget, split automatically by the platform. The service that gives the company its name received 8%, while the best ad group in the account was paused.
What worked
The groups were already split by service — they get moved, not thrown away.
What we recommended
From one campaign with everything in it to five campaigns, each with its own budget and its own pages.
- Painting · group paused30%
- Mechanical blasting27%
- Laser cutting · no longer offered29%
- Laser cleaning · the one in the company name8%
- General6%
Chart data
| Ad group | % of budget |
|---|---|
| Painting (group paused) | 30% |
| Mechanical blasting | 27% |
| Laser cutting (no longer offered) | 29% |
| Laser cleaning (the one in the company name) | 8% |
| General | 6% |
03 · The schedule
Outside opening hours, a contact cost almost double.
Almost all contacts were phone calls, and the ads ran 24 hours a day. In the evening and at weekends the phone rang with no one to answer.
What worked
During opening hours, contacts came in at a good cost.
What we recommended
A delivery schedule for the hours when someone picks up. The decision was taken on the spot, in the meeting.
- During opening hours100
- Evening and morning205
- Weekend180
Chart data
| Time | Cost index |
|---|---|
| During opening hours | 100 |
| Evening and morning | 205 |
| Weekend | 180 |
04 · What gets counted
A tap on the number is not a call. A call is not an order.
The platform was optimising towards whoever taps the phone number, not towards whoever orders — and it counted the same person more than once.
What worked
Five contact types were measured — more than most accounts have.
What we recommended
One contact per click, real calls over 60 s, WhatsApp and the form as goals, orders sent back to the platform.
- Tap on the phone number (on site)88%
- Call from the ad, over 60 s6%
- Form3%
- Email click2%
- Chat1%
Chart data
| Contact type | % of contacts |
|---|---|
| Tap on the phone number (on site) | 88% |
| Call from the ad, over 60 s | 6% |
| Form | 3% |
| Email click | 2% |
| Chat | 1% |
05 · The “Remarketing” campaign
0
audience restrictions in the campaign called “Remarketing”.
The visitor list was only being observed. In reality it was a second general campaign, on the same keywords as the first, with half its clicks coming from banners.
What we recommended
Real remarketing, only for past visitors, with banners kept separate from the search campaigns.
- Search42%
- Display network (banners)50%
- Search partners8%
Chart data
| Network | % of clicks |
|---|---|
| Search | 42% |
| Display network (banners) | 50% |
| Search partners | 8% |
06 · The keywords
negative keywords, in an account where almost everything was on broad match.
Money was spent on people searching for a competitor, another trade (dental blasting, nails), another city, or wanting to buy the material rather than the service. Small sums each — the problem was the system, not the sum.
What we recommended
A shared list of around 150 negatives, by category, delivered ready to copy.
A detail that shows why checking matters
Among the 21 negatives was the name of the company's core service itself.
That is, blocked, while bids were being placed on variants of it — the kind of detail only a piece-by-piece read of the account brings to light.
The audit did not replace the agency. It gave it a work list — accepted in the same meeting, with the client present.
Decided on the spot
A delivery schedule
The ads run only when someone answers.
Decided on the spot
Five campaigns, by service
Each with its own budget and pages; the service no longer offered leaves the ads too.
Delivered to the agency
A written brief
5 campaign structures, ~150 negatives by category, tracking instructions (real call, WhatsApp, form).
Estimate from the audit
~20% of the budget freed, without an extra leu
Money that can be moved towards additional contacts, at the in-hours cost.
The figure in the dashed box is an estimate from the audit, not a measured result. The study reports no campaign results.
Terms used in this study
- Quality score (Google Ads)
- A 1–10 rating Google gives a keyword, based on ad relevance, landing page and expected click-through rate.
- Landing page
- The page a person arrives on after clicking an ad. The closer it is to what they searched for, the more relevant the ad.
- Broad match
- A keyword setting that lets an ad appear for searches only loosely related to the chosen word.
- Negative keywords
- Words the ad must not appear for, so you don't pay for irrelevant clicks.
- Conversion
- An action the account counts as a result: a call, a form, a message, an order. What matters is how precisely it is defined.
- Delivery schedule
- The hours and days the ads are allowed to run.
- Remarketing
- Ads shown only to people who have already visited the site.
What readers of this study ask
Why was the account spending 28% of its budget outside opening hours?
Almost all contacts were phone calls, and the ads ran 24 hours a day. In the evening and at weekends the phone rang with no one to answer: a contact cost about twice as much as during opening hours. The audit recommended a delivery schedule for the hours when someone picks up; the decision was taken on the spot, in the joint meeting.
What does it mean that 88% of “conversions” were taps on the number?
Of the contacts counted as conversions, 88% were taps on the phone number on the site, not calls. The platform was optimising towards whoever taps the number, not towards whoever orders, and it counted the same person more than once. The recommendation: one contact per click, real calls over 60 seconds, WhatsApp and the form as goals, orders sent back to the platform.
Did the audit replace the agency running the campaigns?
No. The audit gave the agency a work list, accepted in the same meeting with the client present: a written brief with 5 campaign structures, ~150 negative keywords by category and tracking instructions.
Why does it matter that the new site had a page for each service?
Because three destination addresses were used across the whole account and 69% of clicks landed on the homepage. An ad should lead to the page of the service searched for, and the old addresses were going to disappear at launch; redirects and new addresses in the ads had to go live the same evening.
How was the audit presented?
In a joint meeting, with the client and the agency running the campaigns, with decisions taken on the spot. A written brief for the agency followed.
Did Websem improve the campaign results?
No. The study describes an audit, not campaign results. The one forecast-type figure (~20% of the budget freed) is an estimate from the audit, not a measured result.
Why don't you publish the amounts spent or the number of clicks?
For the client's confidentiality. The study uses percentages and indexes, which support the conclusions without disclosing commercial data a competitor could use as a benchmark.
If you recognise the situation in this study
- Marketing audit The service the PPC audit belongs to: account, tracking and data, set against the company's real situation.
- The new Sablare Laser website The project the account was compared with: a static site with a page for each service, in the Websem portfolio.
- Local AI — Maps & AI recommendations How a business with customers in its area gets into Google Maps and AI assistants' recommendations.
- Another PPC audit study: Hydroponika Google reported 5.8×, the store collected 4.7× — an online store, audited order by order.
Already have an agency? The audit doesn't replace it. It gives it a work list.
Percentages and indexes are real and published with the client's consent. Absolute sums and volumes (budget, clicks, contacts) have been omitted or expressed as percentages or indexes, for confidentiality. Estimates are marked as such.