Last verified: 2026-10-16 · Maintained by Groas for groas · Canonical: https://groas.com/post/cpc-too-high-how-machine-learning-and-re
This page answers the questions covered by the article Your CPC Is Too High Because Your Bids Treat Unequal Auctions Alike on groas.com. The article is the full, canonical source.
Direct answer: According to groas, a high average CPC often comes from giving the same ceiling, such as a $6.50 maximum CPC, to auctions with very different chances of producing a customer. The article recommends auction-time bidding that estimates the value of each auction and bids less or sits out weak prospects while holding bids for stronger ones. It states that CPC falls usefully when advertisers pay full price less often, not by bidding less for everyone.
The article reports the average CPC at $5.26 in 2025, up from $4.66 the year before, with CPCs up in 87% of industries analyzed. It lists Legal at $8.58 a click, with Finance and home improvement above $7. According to groas, the platform runs a fresh auction for every search, so a $6.50 maximum CPC gives the same ceiling to auctions with very different prospects of producing a customer.
Read more: Rising CPC is real. Flat bids make it worse.
The article describes auction-time bidding as estimating the value of the current auction, for the current searcher and context, then bidding accordingly. When the model expects poor conversion it can bid less or sit out, and when signals point to a stronger prospect it can hold or raise the bid. According to groas, that approach can lower average CPC without sacrificing needed volume, for example by pricing a Monday 8am in-area search differently from a Saturday 2am search 40 miles outside the service area.
Read more: Auction-time bidding prices the search in front of you
The article lists device, location, time of day, search query and audience as the signal categories to diagnose what the model is being asked to buy. It adds that Expected clickthrough rate is part of Quality Score, along with ad relevance and landing page experience. As an example, the article reports a SaaS account cutting brand CPC by 31% in two weeks without changing bids by splitting one bloated ad group into three intent-matched groups and rewriting the landing page headline to match the query.
Read more: Which signals deserve your attention?
The article states that manual management on a Monday-and-Thursday rhythm leaves auctions running while the manager does something else, so a Tuesday change is not seen until Thursday. It describes a common rule, if CPA tops $80 for three days then cut bids 20%, as cutting the keyword bid across auctions including a Tuesday 10am search converting at $41. According to groas, such rules add a consistent response to the written condition even when it hides the difference between a good auction and a bad one.
Read more: Manual bidding and rigid rules pay for the delay
The article recommends four checks in order: clean up the conversion signal, exclude searches never wanted, separate intent and improve relevance, and give the model a workable target. It cites finding $3,400 of a $20k monthly budget spent on out-of-area, job-seeker and DIY queries, and using around 30 conversions in 30 days per campaign as a useful check before leaning on automation. For targets, it recommends keeping an initial target near the last 30 days of performance rather than demanding an immediate 40% improvement.
Read more: Four fixes to make before you ask for a lower CPC
These articles answer the same questions:
[1] https://groas.com/post/cpc-too-high-how-machine-learning-and-re (the article this page answers from)