The short version: keyword tools only show you what people already search for in enough volume to register, which means the exact phrases your best buyers use to describe their problem often show up as zero. The fix isn't a better tool, it's reading your own support tickets, sales emails and the forums your customers hang out in, then writing content around the words they used, not the words a keyword tool approves of.
Why "no volume" doesn't mean "no demand"
I've sat through enough keyword research sessions to know the moment someone's shoulders drop when a phrase shows "0" in Ahrefs or a keyword researcher tool. They assume zero volume means zero interest. It doesn't. It means not enough people have typed that exact string into Google for the tool's sample to catch it. New products, niche B2B problems, and anything emotionally specific ("why does my invoice keep getting rejected by Xero") will almost always show as low or no volume, even when hundreds of your actual prospects are asking a human the same thing every month.
Keyword tools are built on search engine data. Your prospects aren't only searching Google. They're emailing your support team, posting in a Slack community, complaining on a subreddit, or typing a question into a forum thread from three years ago that still gets replies. That's where the real language lives, and it's free to mine.
Start with your own inbox before you touch a tool
Here's where I think most advice gets lazy: people tell you to "read your support tickets" and leave it there. That's not a method, that's a hope. The actual process looks like this:
- Pull the last 200 to 300 support tickets or sales emails from the last 90 days. If you run a small team, this might be your entire helpdesk export; if you're bigger, sample by category (billing, onboarding, cancellations).
- Paste the raw text into a spreadsheet, one row per message.
- Scan for the exact phrase a customer used to describe their problem before they knew the "correct" term for it. Not "integration failure", but "why won't this talk to my calendar".
- Tag each phrase by intent: confused, frustrated, comparing, ready to buy, about to cancel.
- Count repeats. If three or more people phrase the same problem in roughly the same words within a quarter, that's a content opportunity even if a keyword tool says the volume is zero.
This takes a focused afternoon, not a sprint team. I'd budget two to three hours for a first pass on 200 tickets, longer if your support tool doesn't export cleanly.
A worked example
Say you run a small SaaS tool that helps freelancers send client proposals. Your keyword tool shows "proposal follow up email keeps bouncing" at zero volume, and "proposal software" is brutally competitive with agencies already dominating page one. Dead end, according to the tool.
But when you go through 150 support tickets from the last quarter, you notice eleven separate customers describing the same situation in slightly different words: "my proposal email went to their spam", "client says they never got it", "sent but no notification it was opened". None of those exact strings would show volume individually. Together they tell you there's a real, repeated, high-frustration problem around deliverability and tracking.
That becomes a blog post titled something like "Why Your Proposal Emails Aren't Getting Opened (And What To Check First)", written entirely in the language your customers used, not the sanitised industry term. It also becomes a short in-app tooltip and a support macro, because the same phrase will keep showing up. One piece of content now answers a real question, reduces support volume, and ranks for long-tail variations a keyword tool never surfaced because it was too busy being offended by zero search volume.
Forums and communities are a second, richer vein
Reddit threads, niche Facebook groups, industry forums (think Warrior Forum for marketers, or a specific subreddit for your niche), and even Amazon review sections for competitor products are goldmines because people write there with zero filter. They're not trying to rank for anything. They're venting, asking, comparing.
Search inside these communities directly, not through Google. Use the forum's own search bar, or Reddit's search with sort-by-new, and look for the same problem phrased five or six different ways across different threads. Copy the exact sentences into your spreadsheet alongside your support ticket phrases. You'll start to see overlap, and that overlap is your strongest signal, far stronger than any volume number, because it's confirmed by two independent sources: your own customers and strangers who owe you nothing.
One thing worth saying plainly: most of this language will sound clumsy, repetitive, and nowhere near "SEO friendly". That's precisely why it works. Google and readers alike increasingly reward content that matches the messy, specific way a real human describes a problem, because it signals the writer dealt with the problem rather than guessing from a thesaurus. Polished, keyword-optimised phrasing often reads like it was written for a crawler, and increasingly, both readers and search engines can tell.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Turning raw phrases into content that converts
Once you've got 20 to 30 repeated phrases tagged by intent, group them into three buckets:
- Confused/early stage: phrases describing symptoms, not solutions. These become explainer content, like how one of my earlier posts breaks down what a confusing new feature means in plain terms, because people search for clarity before they search for a product.
- Comparing: phrases where customers mention a competitor by name or ask "is X better than Y". These become comparison pages, and they convert disproportionately well because the reader is close to deciding.
- Ready to buy or about to cancel: phrases that show urgency or frustration with their current setup. These feed your sales pages, your FAQ, and your retention emails directly, often word for word.
Write the headline and the opening paragraph using the customer's exact words wherever you can. If eleven people said "never got it" rather than "non-delivery", use "never got it" in your H1. It will feel less professional to you. It will feel exactly right to the next person typing that phrase into a search box or a support chat.
A quick sanity check before you commit hours to this
Not every repeated phrase deserves a full article. Before writing, ask three questions: does this problem affect a meaningful slice of your audience (not just one chatty customer), does solving it publicly reduce support load or move someone closer to buying, and can you answer it better than what currently exists, even if nothing currently exists at all. If you can answer yes to at least two, write it. If you're only mining language to pad out a content calendar, you'll end up with technically accurate posts that nobody needed, which is its own kind of waste.
This whole approach works best as an ongoing habit rather than a one-off project. Set a recurring 60 to 90 minute slot once a month to pull fresh tickets and scan two or three forums. Buyer language shifts as your product and market shift, and the phrases that had zero volume last year might be exactly what a competitor starts ranking for next year if you don't get there first.
Frequently asked questions
What if I don't have enough support tickets to find patterns?
If you've got under 50 tickets a quarter, widen your net to sales call notes, onboarding call recordings, and your own social media comments and DMs. Even 20 to 30 raw customer messages can reveal two or three repeated phrases worth building content around.
Should I still check keyword volume at all if I'm doing this?
Yes, but use it as a secondary check, not a gatekeeper. Once you've identified a phrase from real customers, check broader, related terms to see if there's an adjacent keyword with actual volume you can also target in the same piece.
How long before content based on buyer language ranks?
There's no fixed timeline, it depends on your site's existing authority and the competitiveness of adjacent terms, but because you're often targeting unclaimed long-tail phrasing, you can see indexing and early impressions faster than with contested head terms, sometimes within a few weeks of publishing.
Is this worth doing if I'm not an SEO specialist?
Yes, this is one of the few SEO tasks that doesn't require technical skill, only patience and access to your own customer data. If the volume of work still feels like too much to fit around running your business, it's a reasonable task to hand to an AI consultant for small business who can help you process ticket data at scale and turn it into a content plan.
