Your Google Ads Account Is Too Organized to Get Cheap Leads
Your Google Ads account has 12 campaigns and you are proud of that. One per product line, neat little folders, everything in its place. That tidiness is the reason your cost per lead keeps climbing, and this article defends a blunt position: in 2026, the granular structure that small businesses treat as good hygiene is the single biggest reason their conversions stay expensive. Under Smart Bidding, more campaigns means less control, not more. The fix for a "poorly performing" account is almost always deleting campaigns, not adding budget.
I have watched this flip over 15 years of working in search. The account that looks the most carefully organized is now, more often than not, the one bleeding money. Below is why the old wisdom broke, the exact number that decides whether your account is sliced too thin, and a 30-minute check you can run yourself or hand to your agency as an accountability test.
Every neat little campaign is starving the algorithm that sets your bids
Google's bidding stopped being something you do and became something the machine does. When you turn on Smart Bidding, the algorithm needs to see enough winning outcomes to learn what a good click looks like for you. Heather Brousell, a 20-plus-year practitioner at Single Throw Marketing, put the threshold plainly in Search Engine Land on July 1, 2026: a campaign needs roughly 30 to 50 conversions per month to exit what Google calls the learning phase. Below that, the algorithm is guessing.
Guessing with your money. That is the part nobody translates into dollars for the business owner. WordStream's 2026 benchmarks put the average Google Ads cost per lead across all industries at roughly $66.69, and average search cost per click has climbed to about $2.96 in early 2026, up from $2.64 a year earlier per DigitalApplied. When your campaign never leaves the learning phase, you pay those rising prices while the system bids blind. The waste is not a rounding error anymore.
"Campaign structure determines how Google's machine learning interprets your data, how budget flows across your goals, and whether your data gets collected in one place or scattered across too many campaigns."
Heather Brousell, Single Throw Marketing, in Search Engine Land, July 1, 2026.
Read that again with your own account in mind. Scattered across too many campaigns is not a style preference. It is the mechanism by which a well-meaning owner cuts their own conversion data into pieces too small for the algorithm to use. Ameet Khabra, a 16-year PPC specialist and former Google Partners Ambassador, put the stakes for 2026 plainly in Search Engine Land on June 18: "What has changed is how quickly a poorly architected account can waste budget." The architecture is the problem. Not the bids, not the keywords. The shape of the thing.
Granular structure was correct once, which is exactly why owners still trust it
This is not a case of business owners being careless. They are following advice that used to be right. In the manual-bidding era, you set bids by hand, and slicing an account into narrow campaigns and single-keyword ad groups gave you surgical control over what you paid for each term. The tighter the structure, the more precisely you could tune. That was the whole game, and the advice spread because it worked.
Single-keyword ad groups, the SKAGs, were the height of that thinking. One keyword, one ad group, total control. Agencies built entire practices on them, and plenty still do, because the muscle memory of 2015 is hard to shake. The problem is that the entire logic behind SKAGs assumed a human was setting the bids. Once Google took the bids over, the same fragmentation that gave you control now starves the algorithm of the pooled data it needs. The tactic did not get worse. The ground under it moved.
Brousell's recommended structure reflects the shift: three to five tightly themed ad groups per campaign, not dozens of micro-groups. That is not laziness. It is data pooling. You want related searches feeding the same campaign so conversions accumulate fast enough to clear the threshold. If you have already read my piece on how Smart Bidding quietly moved the budget risk onto you, this is the structural half of that same story. The bidding got smarter, and the account layout that made sense before that change now works against you.
The received wisdom, more campaigns equals more control, is precisely backwards now. Control in 2026 comes from concentration. You give the machine a dense pile of conversion data and it repays you with cheaper leads. You hand it a dozen thin piles and it treats every one as a first day on the job, forever.
The 12-campaign account that could never learn
Brousell's audit contains a case that will feel uncomfortably familiar. An ecommerce account, structured the way any organized owner would structure it: 12 Search campaigns, one per product category. Every campaign averaging 8 to 12 conversions a month. On paper, immaculate. In practice, a machine that never learns.
Do the arithmetic the algorithm does. Each campaign needs 30 to 50 conversions a month to exit learning. Each campaign is delivering 8 to 12. Not one of the 12 campaigns ever crosses the line. Every single one sits permanently in the learning phase, bidding without confidence, spending at those rising 2026 click prices while producing the most expensive leads that account will ever generate. The owner sees 12 tidy campaigns and assumes control. What he actually has is 12 blindfolds.
Now do the arithmetic. Twelve campaigns averaging 8 to 12 conversions each works out to somewhere between 96 and 144 conversions a month in total. That total is my math, not a figure Brousell reports, but it is more than enough, pooled into two or three campaigns, to clear the threshold comfortably and let Smart Bidding actually optimize. The conversions existed. The structure hid them from the algorithm. Consolidation, not more budget, was the entire fix.
"Structure first. Optimization second."
Heather Brousell, Single Throw Marketing, in Search Engine Land, July 1, 2026.
Most owners have the order reversed. They tweak bids, swap ad copy, add negative keywords, all optimization moves, on top of a structure that guarantees none of it can compound. You cannot optimize your way out of a data-fragmentation problem. If the account is sliced too thin, the fixes you are paying for are polishing a car with no engine.
The 30-minute consolidation check you can run without knowing a thing about bidding
You do not need to understand the algorithm. You need one question and a willingness to act on the answer. Here is Brousell's audit logic translated into steps a business owner can do over a coffee, or hand to an agency as a straight accountability test.
Step one: open your account and count conversions per campaign, not per account. The account total does not matter here. In the Campaigns view, go down the Conversions column and write down how many conversions each one produced last month. That single column of numbers is your whole diagnosis.
Step two: ask the one question that exposes everything. Does each campaign get at least 30 conversions a month on its own? If yes, that campaign is fine. Leave it alone. If no, flag it. Any campaign under 30 is running blind and costing you more than it should. You are looking for how many flags you collect.
Step three: group the flagged campaigns by what they actually have in common. If five of your under-30 campaigns are all selling variations of the same thing to the same kind of buyer, they belong together. The test is whether a customer would see them as the same intent. A shopper looking for "running shoes" and one looking for "trail shoes" are close enough to share a campaign. A plumber's emergency-repair leads and their bathroom-remodel leads are not.
Step four: consolidate the genuine matches into single campaigns with three to five themed ad groups each, following Brousell's rule. You are not deleting the keywords or the products. You are pouring several thin puddles of conversion data into one pool deep enough for the algorithm to swim in. Keep separate only what is genuinely a different goal, a different geography with a different budget, or a different profit margin that needs its own target.
Step five: if you use an agency, ask them this exact question and watch how they answer: "why do we have 12 campaigns averaging 10 conversions each, when none of them can exit the learning phase?" A good agency will already have consolidated or will explain the specific separate-goal reason each campaign earns its independence. An agency that responds with a speech about granular control and better organization is describing a 2015 account. Before that conversation, it is worth reading my breakdown of where the average account leaks 25 cents on every dollar, because thin campaign structure is one of the largest leaks on that list.
One more trap inside the structure conversation: Performance Max eating your cheapest wins
While you are looking at structure, look at Performance Max, because it hides inside this same problem. PMax now accounts for roughly 45% of all Google Ads conversions per DigitalApplied's 2026 figures, and that number comes with a vendor-adjacent tilt, so hold it loosely. What is not loose is Brousell's warning about what PMax does to a structure when you let it run unchecked.
"PMax can cannibalize high-intent branded search traffic, inflating costs on terms you would've won cheaply."
Heather Brousell, Single Throw Marketing, in Search Engine Land, July 1, 2026.
Translate cannibalize into money. Someone searches your company name, ready to buy, a click you would have won for pennies through a plain branded Search campaign. PMax, hungry for conversions, jumps in front of that search and charges you a premium for a customer who was already yours. It reports the conversion proudly. You paid a toll on your own front door. Structure fixes this: a dedicated branded Search campaign and the right exclusions keep PMax out of traffic you already own. If you want the deeper version of that argument, my analysis of whether Performance Max is actually driving new sales or just claiming credit covers the incrementality trap in full.
All of this only pays off if you watch the right numbers afterward, and the flattering ones will try to fool you first. The vanity metric to distrust is a rising total conversion count while your cost per lead climbs alongside it. More conversions at a worse price is not progress. It is often PMax cannibalization or a learning-phase account that never stabilized. Watch two numbers after you consolidate: does each surviving campaign now clear 30 conversions a month, and does your blended cost per lead fall. Give the new structure a full quarter before you judge it. That waiting period is my own operating rule, not a figure from Brousell's analysis: a freshly consolidated campaign has to relearn before it settles, and owners who judge it at week three tear down good structure for no reason. If cost per lead has not improved after that quarter of clean structure, the problem is somewhere else, and that is worth knowing too.
Frequently Asked Questions
How many Google Ads campaigns should a small business actually have?
Fewer than you think, and the count is set by conversions, not by how many products or services you sell. The working rule from Heather Brousell's July 2026 Search Engine Land analysis is that each campaign should be able to reach 30 to 50 conversions a month on its own so Smart Bidding can exit the learning phase. If you divide your total monthly conversions by 30, that is roughly the maximum number of campaigns your data can actually support. Many small businesses only have enough conversion volume to justify two or three.
What is the Google Ads learning phase and why does it cost me money?
The learning phase is the period when Google's automated bidding is still gathering data to figure out how to bid for your goal. During it, the system is essentially guessing, so your cost per lead runs higher and less predictably. A campaign needs 30 to 50 conversions a month to exit it. If a campaign never hits that number, it stays in learning permanently and you pay inflated 2026 click prices, around $2.96 per click on average, the entire time.
Should I still use single-keyword ad groups (SKAGs) in 2026?
No. SKAGs made sense when humans set bids manually and needed tight control over each keyword. Under Smart Bidding, that fragmentation splits your conversion data into pieces too small for the algorithm to learn from. The current recommendation from practitioners like Brousell is three to five tightly themed ad groups per campaign, which pools related searches together so conversions accumulate fast enough to clear the learning threshold.
Go count your campaigns and count the conversions in each. If most of them sit under 30 a month, you have found the leak, and it will not close by adding budget. The uncomfortable part is that the account you have been maintaining most carefully is the one costing you the most. Neatness felt like control. In 2026 it is just a dozen blindfolds, and you paid for every one.
