Blocking "Free" and "Cheap" Searches Without Losing Buyers

Every negative keyword starter list on the internet has the same two words at the top: free and cheap. They get pasted into a shared list on account setup, applied to every campaign, and never reviewed again. It feels like hygiene.

It is half right. One of those words is usually pure waste. The other one, in a lot of accounts, is where the margin-conscious buyers live. The starter lists do not distinguish, because a starter list cannot know your prices, your margins or your competitors. Here is how to decide for your own account — and what happens mechanically when you get the match type wrong.

"Free" means the user does not intend to pay. "Cheap" means they do.

That is the whole distinction, and almost every bad decision in this area comes from ignoring it.

Somebody searching "free CRM for small business" is telling you their budget is zero. They may convert in eighteen months when their team grows, but they will not convert on this click, and attribution will never credit the click anyway. Somebody searching "cheap CRM for small business" has accepted that they will pay. They are shopping on price. If you are the cheapest, or you have a $19 tier, that is a qualified visitor.

The same split holds in retail. "free shipping running shoes" is a filter, not a freeloader — that one is arguably a buyer too. "cheap running shoes size 11" is a buyer with a constraint. But "free running shoes giveaway" and "how to get free running shoes" are not customers at all.

So the real rule is not "block free, keep cheap". It is: the word tells you almost nothing; the phrase around it tells you everything. And the only reliable source for what phrases people actually type is your own search terms report, which is also missing a chunk of your terms for privacy reasons — so your read is directional, not complete.

Words that behave like "free"

These signal an intent to not pay at all: free, freeware, gratis, torrent, crack, pirated, no cost, free trial (if you have no trial), free download, free template, free sample, giveaway, win a, sweepstakes.

Most of these are safe to block broadly, with two exceptions worth checking. If you genuinely sell a free tier or run a free trial as the front of your funnel, "free trial" is your best keyword and blocking it is self-harm. And "free shipping" / "free delivery" belong in the "cheap" bucket — those are purchase-intent modifiers wearing a free costume.

Words that behave like "cheap"

cheap, cheapest, budget, affordable, discount, deal, bargain, low cost, under $50, value, second hand, used, refurbished, clearance, sale.

Every one of these is a buyer saying what they will spend. Whether they are your buyer depends on your price. A $4,000 orthodontic clinic should probably block "cheap braces" and "braces payment plan" — not because the person won't pay, but because they won't pay you. A clinic with a $1,800 clear-aligner offer should be bidding on exactly those terms.

That is a pricing decision disguised as a keyword decision, and it should be made by whoever knows the margin, not by whoever is cleaning search terms.

What blocking "cheap" actually costs you

The cost is not theoretical, and it is bigger than people expect, because "cheap" sits in the long tail where volume adds up quietly.

Run this before you block anything. Open the search terms report, last 90 days, filter Search term contains "cheap", and add columns for cost, conversions and conversion value. You will get one of three pictures:

1. Thirty terms, $180 spent, zero conversions. Block the word. Easy.
2. Forty terms, $600 spent, nine conversions at $45 each. Keep the word. Your average CPA elsewhere is probably worse.
3. Sixty terms, $900 spent, three conversions, and all three came from two specific phrases. This is the common one, and it is the one that destroys blunt rules.

In case three, the right move is not to block "cheap". It is to block the thirty losing phrases and keep the two that work. That is more work, which is exactly why people reach for the single-word negative instead. It is also where the money is: a word-level block in case three throws away three conversions to save $700, and the spreadsheet only shows you the $700.

If you want a defensible way to put a number on the saving before you act, calculating wasted spend honestly means subtracting the conversions you will lose, not just totalling the clicks you will stop buying.

The segment nobody counts

There is a second cost that does not show up in any report. "Cheap" and "discount" searchers often convert at a lower order value but a higher rate, because they have already decided to buy and are only choosing where. Block them and your conversion rate drops, your CPA rises, and Smart Bidding — which is learning from whatever data you let through — starts bidding more aggressively on a narrower, more expensive pool.

I cannot give you a published figure for that effect, and anyone who gives you one is guessing. What we see in accounts we audit is that removing a price-modifier word usually drops conversion volume somewhere in the low single-digit percent while the CPA moves the wrong way for a few weeks. Small, but it is a cost, and the person who added the negative never attributes it.

Broad negatives are where this goes wrong fastest

A broad negative keyword cheap blocks any search containing the word "cheap", in any order, regardless of what else is in the query. It does not block close variants — negatives never do. That last part is the trap most people have backwards.

Three things make broad negatives dangerous here:

They catch brand and place names. cheap as a broad negative kills "Cheapside dental clinic" and "Cheap Monday jeans" — a real brand. free kills "Freeport Texas storage units", "Freeman Hospital parking", "sugar free syrup", "gluten free bread delivery", "duty free perfume". That last cluster is the one that bites e-commerce hardest. A broad negative on free in a food or supplements account is a slow, invisible amputation of your best queries.

They do not respect your intent across match types. Broad negatives behave unpredictably once you mix them with phrase and exact keywords, and the interaction is not documented in a way you can plan around. We cover the mechanics in why negative match types don't work like positives — the short version is that a broad negative is blunter than the broad positive match you are used to, not more flexible.

They are invisible once added. A phrase negative cheap tickets reads clearly in the list six months later. A broad cheap reads like a policy. Nobody questions a policy.

The safer default for price words is a phrase negative on the losing phrase, not a broad negative on the word. "cheap knockoff", "cheap alternative to", "free pdf", "free download". Longer, more of them, far less collateral damage. If your list is getting long, that is usually fine — the real limits are discussed in how many negative keywords are too many, and the answer is not the number you fear.

The one place broad is right

If a word can only ever appear in a wasted query in your vertical, broad is correct and simpler. A commercial roofing contractor can block salary, jobs, torrent and wikipedia broadly and sleep fine. The test is: can you imagine a single paying customer typing this word? If you hesitate for more than two seconds, use phrase.

Performance Max makes the free/cheap decision harder, not easier

Performance Max will serve on "free" and "cheap" queries without telling you much about it, and your tools for fixing that are narrower than in Search.

Three constraints you are working under:

• A Performance Max campaign accepts up to 10,000 campaign-level negative keywords. That is generous.
• It does not accept shared negative keyword lists. Your carefully maintained account-wide "free/cheap" list does not apply. You push the same terms into each PMax campaign separately, or you do not block them at all.
• Search terms for Search, Shopping and Performance Max all land in campaign_search_term_view. Display, Video and App report nothing — so a PMax campaign spending heavily on Display inventory can soak up freeloader traffic that never appears as a search term anywhere.

That third point matters for this topic specifically. Free-seeker traffic is cheap traffic, and cheap traffic flows to the cheapest inventory. The clicks you are most worried about are partly in the places you cannot see. What the PMax search terms report actually shows goes through what is and is not in there, and where Performance Max wasted spend hides covers the rest of it.

The practical consequence: in PMax, be more willing to block "free" and less willing to block "cheap". You have weaker feedback on what the price-modifier queries are doing, and blocking a demand segment you cannot measure is how a PMax campaign quietly stops finding volume. If you are sitting on thousands of negatives per campaign, the ceiling and how to spend it are covered in negative keywords in Performance Max and the 10,000 cap.

Seven decisions to make before touching the list

Answer these for your own account. They take fifteen minutes and they replace every generic starter list.

1. Do you sell anything free? A trial, a freemium tier, a free sample, a free quote, a free consultation. If yes, "free" is not a negative — it is a landing page. Block free download and free template and keep free quote. Local service businesses get this wrong constantly; "free estimate roof repair" is one of the highest-intent queries in the trade. More patterns in local business wasted spend.

2. Where do you sit on price in your category? Top third, bottom third, middle. Bottom third: never block price modifiers, bid on them. Top third: block the hardest ones (cheapest, cheap, budget) and keep the softer ones (deal, sale, discount), because even premium buyers look for a discount code.

3. Do you sell used, refurbished or clearance stock? If you clear last season's inventory, "cheap", "used", "second hand" and "clearance" are your inventory pages. Blocking them orphans real SKUs. Common failure in retail accounts — more in e-commerce search terms that drain budget.

4. Is the word ambiguous in your language or market? "Free" in English is a minefield: sugar free, gluten free, hands free, duty free, freelance, Freeport, Freetown, free range. List the collisions before you block.

5. Does the word appear in a competitor's name? Check. Then decide separately whether you want competitor traffic at all — that is its own question, handled in competitor searches, block or keep.

6. What did the word cost you in the last 90 days? Not opinion. The filtered report. If "cheap" cost $40 total, you are arguing about nothing; spend the time on the terms that cost $400.

7. Who reverses this if it is wrong? If the answer is "nobody, because nobody will notice", do not add the broad negative. Add phrase negatives you can read and undo later. Undoing negative keywords is easy in principle and awkward in practice once the list is a year old and nobody remembers why.

How to review price-modifier negatives on a schedule

Once a quarter, pull every negative in the account containing a price word and ask whether it is still true. Prices change, competitors change, and a negative added when you were the premium option is a liability after you launch a budget line.

The cheapest version of this review:

• Export all negative keywords, account lists and campaign level both.
• Filter for free, cheap, discount, budget, afford, low cost, bargain, used, second hand, cost.
• For each one, ask: would a customer we want ever type this? If you cannot answer without checking the price list, you have found the ones worth reviewing.
• Remove the broad ones. Replace with phrase negatives on the specific losing queries.

Then watch impressions for two weeks. If a word was doing real damage, nothing happens when you remove it. If impressions and conversions both tick up, you had been blocking buyers. The symptoms of over-blocking are listed in six signs your negative keywords are hurting performance — low impression share with budget left over is the clearest one.

Agencies managing twenty accounts cannot do this manually per account per quarter; the structural answer is a shared layer for the universally-dead words and a per-account layer for the price words. That split is the main idea in managing negative keywords across client accounts.

A note on review frequency

Price-modifier terms do not need daily attention. The rest of your search terms do — how often you should check search terms argues for far more often than most accounts manage. The distinction: a genuinely irrelevant term ("plumber salary" in a plumbing account) is wrong today and wrong forever. A price-modifier term is wrong only relative to your current prices, so it belongs on a slower, more deliberate cycle with someone who knows the business in the room.

Where PPCleaner draws the line on price words

PPCleaner reads your search terms every night and adds negatives for the terms that were never going to convert. It is deliberately conservative about price modifiers, for the reason this whole article is about: cheap is a commercial judgement, not a relevance judgement.

What that means concretely:

• It never blocks a term that converted in the last 90 days. If "cheap running shoes mens" brought a sale, it stays, full stop.
• It judges terms, not words. It will add "free pdf download" as a phrase negative after it has spent money with nothing to show. It will not add a broad free across your account because a few "free" queries lost money.
• It writes negatives at the level that works: campaign-level for Performance Max, since PMax takes no shared lists, and the usual structures for Search and Standard Shopping, where Shopping negatives behave differently enough to matter.
• Every block is reversible in one click for 90 days. If it blocked something that was quietly producing revenue, you take it back and the term is live again.

What it cannot do: it does not know your margin. If you sell at a premium and you want every budget-shopper query gone on principle — even the ones that convert, because they convert at a loss — that is a decision you make and enter yourself. PPCleaner optimises against conversions, not against profit, because it cannot see your profit. On a product where the cheap-segment orders lose you $8 each, a conversion-based system will protect exactly the traffic you want killed.

It also never touches budgets, bids, ads or keywords. If your real problem is that you are bidding on "cheap X" with a broad match keyword and Google is matching it to everything adjacent, negatives are a bandage. The underlying question is covered in are broad match keywords worth it, and the diagnostic order for an account spending with nothing coming back is in check this order.

The short version of the rule

Block the words that mean "I will not pay". Keep, test and price the words that mean "I will pay less".

For most accounts that comes out as: broad negatives on torrent, crack, freeware, giveaway. Phrase negatives on "free download", "free template", "free pdf", "how to make", "diy". And nothing at all on cheap, affordable, discount or budget until you have looked at the last 90 days of data for those words specifically and can say out loud what they cost and what they returned.

If that feels like less blocking than you expected, good. The accounts that waste the most money are not usually the ones with too few negatives — they are the ones with a hundred negatives added in one afternoon in 2021 and nobody left who can explain them. A quick pass through the one-hour audit checklist will usually turn up at least one broad price negative that has been quietly suppressing good traffic for a year.