Why I leave about $50 a day unspent in my own Google Ads account

Between 21 and 25 August my own Google Ads campaign spent $51.04 across 8 clicks against a daily budget of $64.52. Read in the campaigns view with both impression-share columns showing, that window lost 0.0% of impression share to budget and 75.1% to rank. The 11 to 17 August window on the same two columns read 84.3% and 12.5%. Roughly $50 a day now goes unspent, I know why, and I’m leaving it there.

Three changes, in order, and then the flip

I made all three of these myself, in this sequence, and only the last one looks like an experiment.

On 14 August I culled the keyword set. Click volume fell, which was the point. An audit I ran a week later read the fall as an over-correction and tried to re-enable four of the paused keywords. I reverted all of it the same day.

On 17 August I wrote a bid plan against a daily budget of $14.61. Nearly every enabled keyword went to a $6.50 cap. I set it low on purpose, so one $30 click couldn’t eat a whole day at that budget.

On 21 August I moved the monthly target to $1,000 and the daily budget went to $64.52. My pacing tool, AdPulse, runs at a 2x limit, so that figure is exactly 2 x ($1,000 / 31). It’s a ceiling rather than a computed catch-up number. I never re-tuned the bids underneath it, so the campaign spent the following week carrying a $14.61-a-day bid plan on a $64.52-a-day budget.

Here’s what the weekly review pulled on 25 August, three windows from the campaigns view with Search Lost IS (budget) and Search Lost IS (rank) added:

Window	Spend	Clicks	Leads	Lost IS, budget	Lost IS, rank
11 to 17 August	$183.01	29	0	84.3%	12.5%
18 to 24 August	$75.80	11	0	22.6%	60.4%
21 to 25 August	$51.04	8	0	0.0%	75.1%

One mechanical note before that table gets over-read. Impression share, lost to budget and lost to rank add up to 100%, so the 12.5% in the first row couldn’t have exceeded 15.7% while the campaign sat out of money 84.3% of the time. What the third row shows is the rank loss with the budget ceiling taken off it. That’s a clearer view of the same auction, not evidence of a harder one.

The same weekly review puts the campaign at $11.71 a day on a trailing seven-day basis against that $64.52 budget. The budget increase revealed what the cull and the bid cap had already done to the account, and it took away the last excuse for the spend. Caps of $6.50 sit under first-page bid estimates of $13 to $32. The campaign can’t enter the auctions it wants, however much money is parked behind it. So the question stopped being why spend stalled and became whether clearing the floor is worth doing. I decided it isn’t. The reason is arithmetic rather than nerve.

The arithmetic that stopped the bid raise

August closed the 25th at $534.28 across 75 clicks with one genuine lead, which arrived on 9 August and reconciles against a real generate_lead event in my own analytics. One lead per 75 clicks is a 1.33% lead rate. That’s the only conversion figure this account owns, it comes from the same account over the same month as the spend above, and everything below stands on it.

Four steps, and I would run them in this order on any account where the rank column has taken over:

1. Take the two lost-impression-share columns apart before touching anything. Lost to budget and lost to rank point at opposite fixes. Mine moved from 84.3% budget in mid-August to 0.0% budget and 75.1% rank across the 21st to the 25th, which is a different account from the one the bid plan was written for.
2. Pull the first-page bid estimate, not the market CPC. Mine runs $13 to $32 across the enabled keyword set against a market CPC nearer $7. Google computes that estimate from the keyword’s own Quality Score and from what competitors are bidding, so the gap between those two numbers is partly the price of my own account, and it’s the number a bid has to clear.
3. Multiply the floor-clearing bid by clicks per lead, and read the result as a ceiling. At a 1.33% lead rate that’s 75 clicks per lead. The weekly review put floor-clearing bids at $16 to $23, which lands cost per lead between $1,200 and $1,700. That’s arithmetic off the bid cap rather than a realised CPC. You pay a cent over the next bidder rather than your own cap, so treat $1,700 as the top of the band.
4. Divide by the close rate, because cost per lead isn’t the number that decides this. Cost per booked job is cost per lead divided by close rate, which is the formula I’ve already published for home-services accounts and it applies to mine on the same terms. At the ceiling, an offer with a $5,000 floor needs a 24% to 34% close rate to return the offer price before I’ve done a minute of the work. My account has produced one lead. There’s no close rate here to divide by.

So the caps stay at $6.50 and about $50 a day goes unspent. The raise asks me to spend somewhere near $2,000 buying clicks to discover a close rate on a channel that’s produced a single lead, and I would rather put that $2,000 of attention on the thing the account is measurably failing at.

The lever is the landing page, and the account says so twice

The Quality Score components here are unambiguous. Every keyword carrying a score reads post_click_quality_score: BELOW_AVERAGE alongside creative_quality_score: ABOVE_AVERAGE, which are the API’s names for landing page experience and ad relevance. Google describes Quality Score itself as a diagnostic rather than an auction input, and the auction reads ad quality in real time through Ad Rank instead. So the ads are being read as relevant and the page behind them isn’t. The first-page estimate moves with the keyword’s Quality Score, which makes a below-average post-click read a plausible part of why my estimates sit at $13 to $32 against a market CPC nearer $7. No bid change touches it.

Read by landing page through 25 August, /ecommerce-website-design/ has taken 43 clicks and $272.16 for zero leads. /shopify-store-design/ has taken 29 clicks and $243.07 and produced the one lead. More money has gone into the page that’s never converted than into the page that has.

Both of those are repository changes with an auto-deploy behind them. Neither is an ads write. That took me longest to accept, because the account is the surface I’m fluent in and the page is the surface that decides the price.

The kill gate had to be rewritten too

I set the original gate as volume: by 31 October, roughly $2,500 and roughly 375 cumulative clicks, and under 5 qualified leads means the channel stops. On 25 August I worked out that it could only ever return “kill.” At the trailing seven-day rate the campaign lands nearer $1,320 and 190 clicks by 31 October, and five leads on 190 clicks needs a 2.6% lead rate against a measured 1.33%. The gate was going to fail on volume rather than on whether paid search works here. It was measuring the wrong thing.

It now reads as a lead rate at about 200 clicks. Under 2% and paid search doesn’t clear an offer at this price, so the money moves to outbound. Between 2 and 4% and the conversion rate gets fixed before a dollar of budget moves. Over 4% and it scales. The spoke version of this decision, written as the question people search, is should I raise bids when the budget is not spending.

What I did not claim

This account produced one lead in the window I’m describing, on 9 August, and zero in the sixteen days after it. This isn’t a success story and I’m not writing it as one. There’s no before-and-after here and no lift to point at, only a decision with its arithmetic on the page.

I’m also not claiming that budget increases never move spend. Mine didn’t move because I’d already capped the bids under the first-page estimates four days earlier, which is a specific condition rather than a general rule. On a campaign losing impression share to budget, more budget is exactly the right lever, and my own 11 to 17 August window at 84.3% lost to budget is what that looks like.

The 1.33% lead rate rests on one lead. I’ve never measured a close rate on this channel at all. That’s enough to decline a bid raise, because declining costs nothing and the raise costs about $2,000. It’s nowhere near enough to be a benchmark for anyone else’s account, and the $1,200 to $1,700 figure is a ceiling computed off my bid cap rather than a bill I’ve paid.

Finally, I haven’t proven the landing page causes the low lead rate. What I have is a below-average post-click score on every scored keyword, an above-average creative score on the same keywords, and one page absorbing more spend than the other while producing nothing. That’s where I’m looking next, not something I’ve already fixed.

One thing to do next

Open your campaigns view, add Search Lost IS (budget) and Search Lost IS (rank), and read the two columns for the last seven days against the seven before that. If rank has taken over from budget, the next number you need is the first-page bid estimate, and the number after that is your own measured close rate. If you want somebody to read the account settings underneath all of that, the ten I check first are in the free Google Ads Setup Audit.

Keep going

If this hit, the next two pieces in the same universe:

• The field your Google Ads audit is reading is the wrong one. Same account, twelve days earlier, and a wrong finding I published against myself.
• Max Clicks, 0 conversions. The opposite failure: a bid strategy buying all the volume it can, and none of it human.

Free PDF: The 25-page Google Ads Setup Audit. No email gate.