← Back to blog

Strategy · November 10, 2026 · 8 min read

Zero Search Volume Keywords: Are They Worth Writing?

A zero is a reporting floor, not a verdict on demand. Two of the three things that produce one are worth writing for, and a ten-minute check tells them apart.

By FluxWriter

Zero search volume keywords are worth writing in two of the three situations that produce a zero, and a waste of a post in the third. The zero itself tells you very little, because it is a statement about your tool's sample rather than about human behaviour. What separates the three cases is a check you can run in about ten minutes, before you commit a writer or a budget. This post covers the three cases, that check, and the arithmetic that decides whether a portfolio of these posts is worth building at all.

A Zero Is A Reporting Floor, Not A Measurement

A keyword tool does not count searches. It models them. Most tools blend clickstream data bought from panels of real browsers with Google's advertiser-facing figures, then extrapolate to a national estimate. Google Keyword Planner, which sits underneath much of this industry, reports in ranges rather than exact counts and suppresses the smallest ones. Every tool built on top of that has a floor below which it stops reporting and prints a dash.

So a zero is a sentence about the tool. It says the phrase did not appear often enough in the sample to survive rounding. It does not say nobody types it. Those are different claims, and the gap between them is where most of the long tail lives. Google has said publicly that a meaningful share of the queries it handles each day are ones it has never seen before, which means a large amount of genuine demand exists in phrasings no historical dataset can hold.

There is one dataset without that problem, and you already own it. Google Search Console reports impressions for your own pages, including queries that get a handful of impressions a month and would show as a dash in any third-party tool. Some rare queries are withheld for privacy, so the picture is incomplete, but what it does show is measured rather than modelled. Anyone whose keyword research consists entirely of a third-party estimate is arguing about a model when there is a partial record of reality sitting in their own account.

Case One Is The Term Your Tool Cannot See Yet

The first case is a blind spot. The phrase is real, people type it, and the tool has not caught up. This happens most often with anything recent: a product released in the past year, a feature name, a pricing change, a regulation that shifted, an error message a new software version started throwing. Clickstream models need months of accumulated behaviour to produce a stable estimate, so recency and low volume look identical from the outside.

It also happens with length. The longer and more conversational the phrase, the more likely it is that real search traffic is spread across dozens of near-identical wordings, none of which individually clears the reporting floor. The demand is there in aggregate. The tool simply has no row for it, because the tool indexes strings rather than intents.

These are usually the best posts you will write. The competition for them is close to nothing, because everyone else filtered them out of the same export you did. The trade is that you cannot forecast the return before publishing, so you are betting on the topic being sensible rather than on a number.

Case Two Is The Vocabulary Only Your Customers Use

The second case is the phrase your buyers say and the wider market does not. Every industry has them. The manufacturer's catalogue name for a part, the internal shorthand a procurement team uses, the term a specific piece of software prints on screen, the way a regulated sector describes a document that has a plainer name everywhere else. These terms have small, specific, high-intent audiences, and no keyword tool will ever price them properly.

The corpus for finding them is not a keyword tool at all. It is your support inbox, your sales call notes, your live chat transcripts and your helpdesk tickets. Search those for phrases that recur, then check whether you have a page that uses the same words your customers used. These pages tend to convert at a far higher rate than head terms do, because a person typing a part number or an internal term is already deep in a specific problem rather than browsing.

This case is also where being the only usable answer matters most. When a page is the single clear source on a narrow phrasing, it has a better chance of being the source that gets quoted, linked or summarised elsewhere. That is an advantage worth holding, though it is not one you can measure in advance.

Case Three Is Demand That Was Never There

The third case is the one to drop. These are phrases produced by a generator rather than by a person: a head term crossed with a modifier crossed with a location, permutations that look like keywords and were never typed by anybody. They also turn up in translated lists, where a phrase that is common in one language becomes a construction nobody uses in another.

The tell is in the results page. Google returns something for every string, so the presence of ten results proves nothing. What matters is whether those results are about the phrase or about the nearest common query the engine substituted. If the top ten all address a broader term and nothing addresses the exact one, Google has decided the exact phrasing has no distinct intent behind it, and a post that targets it will be judged against the broader term it cannot win.

Writing these is not harmful in itself. It is simply the wrong allocation. A page nobody searches for and nobody links to is a page you paid for and will maintain, and a portfolio full of them makes a site look padded rather than deep.

Five Checks Separate The Cases Before You Commit

Start by searching the exact phrase and reading the top ten results with one question in mind: are these pages about this phrase, or about something broader. Then check whether Google's autocomplete completes the phrase and whether People Also Ask returns questions that use its vocabulary. Autocomplete is built from queries people actually submitted, so a suggestion is stronger evidence of demand than any modelled volume figure.

Next, open Search Console, filter queries by the distinctive word in the phrase, and look for impressions. Even two or three a month tells you the phrase exists and that Google already associates your site with it. Then search your own support and sales records for the same words. Finally, check the parent topic: if the head term has measurable volume and your phrase is a specific version of it, the tail underneath it is almost certainly real even when it cannot be measured.

As a rough rule, we treat two positive signals out of those five as enough to write, and a phrase that fails all five as case three. The checks take about ten minutes per phrase once you are used to them, which is the point. The decision has to be cheaper than the post.

The Arithmetic Only Works If You Publish Enough Of Them

One zero-volume post is not a strategy. The returns on any individual page are small and unpredictable, so the model only works as a portfolio, where a handful of pages carry the result and the rest cost you little. The numbers below are illustrative rather than measured. They assume half of these posts eventually pick up any search traffic at all, that the ones that do settle around five visits a month, and a freelance rate of $250 a post.

Two things fall out of that table. The first is that six posts a year cannot produce a result worth measuring under these assumptions, no matter how well chosen the phrases are. The second is that cost per visit stays flat at roughly $8 across every row, because the input is priced per post. Scale raises the total and does nothing to the unit cost. The only lever that moves the unit cost is the price of producing a post.

That is why this strategy reads as impossible to most B2B and niche sellers. They are told SEO cannot work for them because their tools return dashes, when the real constraint is throughput. Their traffic has to come from many specific pages rather than a few broad ones, and most of them are buying content at a rate that makes the arithmetic fail before the keyword choice matters. Fix the cost per post or the number of posts, and the same phrase list produces a different answer.

FAQ

Does Google treat a page about an unsearched term as thin content?

Search volume is not the test. Google's published guidance is about whether a page is useful to the person reading it, and a short, specific, genuinely helpful answer is not penalised for having a small audience. The real risk in this strategy is publishing near-duplicate pages that differ only by a swapped modifier, which is a different problem from low volume.

How long should I wait before deciding a zero-volume post failed?

As a rough rule, give it six months and judge it on impressions rather than clicks. Impressions climbing while clicks stay flat means the page is being seen and the title or the match is wrong. Impressions still at zero after six months usually means case three, or that the page is not indexed at all, which is worth checking before you conclude anything about demand.

Should these terms get one long page or several short ones?

If the phrases are wordings of the same question with the same answer, put them on one page and use the variants naturally in the body. If each phrase has a genuinely different answer or a different buyer behind it, separate pages will serve both readers better. The deciding question is whether one page can satisfy both searchers without padding.

The Practical Takeaway

Take the list of dashes your tool returned and run the five checks on the twenty that look most like something a customer would say, then write the five that pass most cleanly and leave the rest. Give them six months and watch impressions, not rankings. The single first action is to open Search Console, set the date range to the past twelve months, and export every query with fewer than ten impressions: that list is real measured demand your tool reports as zero, and it is the highest-confidence starting set you will find anywhere.

When a phrase that passed the checks goes to FluxWriter, put the customer's exact wording in the topic angle, because that is what it builds the post's focus keyword from, and in case two the exact wording is the whole point.

← All posts