JAN-PRO Commercial Cleaning
Google Ads Performance
Week Two Report
29 July 2026

Week two told us exactly what was holding us back. We fixed it today.

21–28 July 2026 · JAN-PRO Commercial Cleaning Aotearoa

The short version

Week two was the week the account finally had enough traffic to diagnose properly — and the diagnosis was clear. We were being outranked, not outspent. Google was showing our ads on roughly four in ten of the searches we should have been winning, and the reason was a quality rating we could see, measure and fix.

We found the cause, and we rebuilt around it today: five new region-specific landing pages, a much simpler booking journey, and a keyword list restored after Google's own automation had quietly trimmed it. That's the whole picture below, in plain English.

Week two by the numbers

1,201
Times your ad was seen
week one: 178
60
Clicks to your site
week one: 13
5.0%
Click-through rate
healthy for this market
$394
Spend for the week
week one: $49

The step up was deliberate. We raised bids on 27 July to buy our way into more auctions and get a readable sample — you can't diagnose an account on 178 impressions. It worked: week two produced almost seven times the visibility of week one, and with it, the data we needed.

RegionAd shownClicksSpendShare of available searches
Auckland67032$210.7840.6%
Christchurch20013$88.4047.7%
Wellington1598$56.5942.0%
Waikato1387$38.2557.8%
Bay of Plenty340$0.0018.0%

Bay of Plenty is the outlier and we're on it separately — it has budget allocated but has not yet found traffic. That's a specific piece of work for the coming fortnight, not a mystery we're ignoring.

What we learned

We weren't losing on budget. We were losing on rank.

When your ad doesn't show, Google tells you why. For JAN-PRO the answer was consistent across every region: roughly half the time we didn't appear, it was because we were outranked — beaten to the slot by another advertiser. Budget almost never ran out.

That's genuinely good news. Budget problems cost money to solve. Rank problems are usually a quality problem, and quality is something we control.

Google grades every keyword out of ten. Ours were scoring one to three.

Google's Quality Score is built from three components. Two of ours were fine or improving. The third was the anchor:

ComponentWhat Google asksOur rating
Ad relevanceDoes the ad match the search?Average to above average
Expected click rateWill people click it?Below average
Landing page experienceDoes the page deliver what the ad promised?Below average — every keyword, every region

Landing page experience was the only component rated poorly everywhere. So we went and read the page the way Google reads it — and the problem was immediately obvious.

Our ads and our landing page were talking about different things

The ads were about office cleaning in Auckland. The page was a well-written argument about trust and reliability — genuinely good writing — that happened never to use the words “office cleaning” at all, and never named a single city.

So someone searching “office cleaning company Auckland” clicked an ad about office cleaning in Auckland, landed on a page that mentioned neither, and Google marked us down for it. Every time.

Google's automation had been quietly removing our keywords

A separate find, and a useful one. Google offers an “auto-apply” feature that makes changes to accounts automatically. On two occasions — 16 and 25 July — it decided two of our best keywords were redundant and deleted them: office cleaning and commercial cleaners, across all five regions.

They weren't redundant. They were two of the highest-intent terms in the build. We've restored them and switched the automation off, so it can't happen a third time.

What we changed today

1
Five region-specific landing pages, one per franchise area. Auckland, Wellington, Christchurch, Hamilton and Tauranga now each have their own page that names the city, names the suburbs we cover, and leads with the service people actually searched for. No more one-page-fits-all.
2
The wording now matches what people type. “Office cleaning” went from appearing zero times on the page to appearing throughout, alongside medical, school, warehouse and aged-care cleaning — and the local suburb list came straight from the real searches we'd seen, like Albany, the North Shore and Rotorua. The strong writing stayed; it just now says what the service is.
3
Booking is now one tap to start. Previously you chose your type of site and then had to find and press a separate button to continue. Now choosing your site type is the action — it moves you straight on. One less thing to do, one less place to stall.
4
Each step now lands where it should. Moving through the form used to jump you back to the top of the page, so you lost your place. It now holds you at the form, where you're actually working.
5
We ask for less. Your region is already known from the page you landed on, so it's filled in for you. Email is now optional — for a call-back, a name and a number is all we genuinely need. The button now says “Book my free discovery call” rather than the generic wording it had.
6
We stopped paying to appear against competitors' names. Around one in five of the searches we were showing against were people looking for a named competitor — and they clicked at a third of our normal rate. That drags the quality rating down for everyone else. We've excluded seventeen more competitor brands.
7
Our first exact-match keywords. City-specific searches like “commercial cleaning Christchurch” were taking our most expensive clicks through a looser match. Those now have precise keywords of their own.

What to expect next

Give the quality rating two to three weeks

This is the one thing worth being patient about. Google re-assesses landing page quality slowly — it doesn't re-grade the day you publish. We would expect to see the quality scores start to move through mid-August, and impression share improve behind them. A flat reading this week wouldn't mean it hasn't worked.

On month one, straight

July closed at $443 of the $2,500 measurement budget. We under-spent because we were being outranked, not because we were holding back — and we'd rather tell you that plainly than spend the balance on clicks we already knew were underperforming.

Month one was always the measurement window, and it did its job: we now know precisely what was wrong and it's been fixed. August is the month that should produce a real cost per enquiry, and that's the number we'll be reporting against from here.

Happy to walk through any of this on a call.