The keyword research missed the money terms

There is a kind of SEO failure that is almost impossible to see from the inside, because nothing looks broken. The site is live, the reports turn up on time, the traffic line points up and to the right, and the phone rings about as often as it did a year ago. When I dig into an account like that, the problem is usually not the work. It is the target. The keyword list was built to look respectable in a report rather than to bring in customers, and everything downstream has been aimed at the wrong thing ever since.

This is the first failure in my series on why SEO agencies fail, and it is first for a reason. Every other part of the job depends on the keyword list being right. Get it wrong and you can do everything else well and still lose.

What a money term actually is

A money term is the phrase a person types about thirty seconds before they ring somebody. It is not clever, not broad, and rarely impressive on a slide. Usually it is a service and a place, or a problem and a place, typed by someone who has decided to spend money and is now choosing who gets it.

Take a plumber in Papakura. The money terms are blocked drain Papakura, emergency plumber Papakura, hot water cylinder replacement Papakura. Somebody typing that has water on the floor and will ring one of the first three results.

What ends up on the report instead is the company name, the word plumbing on its own, and a few question-shaped phrases like how to unblock a drain. But somebody searching that is trying not to pay a plumber. Ranking first for it gets you a visitor who leaves the moment they find the coat hanger trick.

The keyword that makes you money is usually the one that looks least impressive in a report. That is the whole problem in one sentence.

Why the wrong terms end up on the list

This is rarely dishonesty. It is mostly the output of a workflow nobody questioned, and four forces all push the same way: away from your revenue.

Volume is the number that looks good

Keyword tools sort everything by monthly search volume, so the first screen anyone sees puts the biggest, vaguest terms at the top. A broad industry word might show thousands of searches a month; the suburb phrase that actually produces a booking might show a couple of dozen. If you are building a list to present to a client paying real money every month, the pull towards the big number is enormous. It fills the page and makes the opportunity look like it justifies the fee. Nobody has to be cynical for that to happen.

Commercial terms look too small to bother with

The flip side of the same problem. A phrase with thirty searches a month reads like a rounding error, so it gets filtered out before anyone asks what those thirty searches are worth. Thirty searches from people with a burst pipe is a different asset entirely to three thousand from students and tyre-kickers. Volume tells you how many people typed something, not whether any of them wanted to buy. Those figures are estimates anyway, and plenty of valuable local phrases show as zero volume while still bringing in work.

Brand terms inflate the win column

If your business name is on the tracked list, you will rank first for it, because you are the only business with that name. It goes green immediately and stays green forever, and line one of the report is a first-place ranking.

There is an innocent explanation, and a decent one: tracking your own name is a health check, because if you ever stop ranking first for it something serious has gone wrong. The failure is counting it as a win and letting it carry the average position figure reported as progress. Brand and non-brand belong in separate sections.

Nobody asked the person who answers the phone

This is the big one. Most keyword lists come out of a tool, seeded from the client's existing website or a generic industry starting point. That produces the language the industry uses about itself, not the language customers use. A clinic calls it musculoskeletal assessment; the person in pain types back pain physio near me. An electrician's site says switchboard upgrades; the customer types fuse box keeps tripping.

An agency working across twenty accounts does not know that half your work comes from three suburbs and none from the fourth. That knowledge lives with you and whoever answers your phone. If nobody went and got it, the list was always a guess.

Informational, commercial, transactional

One piece of jargon is worth learning, because it is the whole game for a local business. Every search has an intent behind it, and they sort into three groups.

• Informational. The person wants to know something: how to unblock a drain, is sciatica permanent, what does rewiring a house involve. Often they are researching so they do not have to pay anybody. Useful for trust, but it will not fill next week's diary.
• Commercial. The person is comparing options: best physio in Hamilton, plumber reviews Tauranga, how much does a bathroom renovation cost NZ. They have accepted they will spend money and are working out with whom. A lot of the real opportunity sits here, and it is the group most often missing entirely.
• Transactional. The person is buying now: emergency electrician Christchurch, dentist open Saturday Wellington, drain unblocking near me. Lowest volume, highest value, shortest distance between the search and the invoice.

A healthy local keyword list leans heavily on the bottom two, with informational terms as a deliberate minority chosen because they lead somewhere. The broken lists I inherit are usually the exact inverse: almost all informational, a couple of enormous single-word terms nobody will ever win, and the brand name holding up the average. That said, if you sell something people research for months, informational content does real work - you need to be in the conversation long before the decision. For an urgent local service it sits on top of a foundation that has to exist first.

Why near me and suburb terms matter

Two things make local phrasing far more valuable than the raw numbers suggest. The first is intent. Adding a place name or the words near me is what somebody does when they intend to physically deal with a business. Nobody types plumber Onehunga while idly browsing. Location is the buying signal, and the most reliable filter for separating customers from readers.

The second is winnability. The huge city-level term is contested by every established competitor plus the directories, and those results are decided largely by authority - how much of the web already trusts each site. That is slow, expensive ground, and exactly the situation I describe in why a page gets stuck at number three. The suburb term is a much smaller field. You can win it this quarter, and a stack of small wins across your service area usually adds up to more work than the trophy term.

Near me searches are widely misunderstood. You do not rank for them by putting the words near me on the page. Google resolves that search using the searcher's location and serves whatever is close, relevant and well regarded nearby. So near me performance is mostly a Google Business Profile job: correct category, correct address, real reviews, service areas that match reality. If your list has near me terms on it but nobody has touched your profile in a year, list and work are not connected.

The ten-minute self-check

No tool and no SEO knowledge required. You need your last report and ten minutes.

1. Write down the five phrases you would type if you were the customer, in the words a customer would use, with your city or suburb on the end. Do not think like a marketer. Think like the person standing in a kitchen with a blocked drain.
2. Sanity-check them. Ask whoever answers your phone what people say when they ring, and read the last twenty enquiries.
3. Open the keyword list on your last report and count how many of your five appear. Exact or very close counts. Vaguely similar does not.
4. Cross out the brand terms. Every line with your business name in it is a health check, not a result.
5. Sort the rest into the three intents. How many are informational questions, how many are somebody comparing, and how many are somebody ready to buy today?
6. Read the verdict. If three of your five are missing, the strategy is aimed somewhere other than your revenue. Anything off that list is not being worked on and not being measured, so nobody would notice if it went backwards.

How to build the money-term list yourself

You already hold better raw material than any tool can generate. This takes an afternoon.

1. Mine your own enquiries. Go through three months of calls, emails and forms and write down the words people used. Not your words. Theirs. Best keyword research there is, and it costs nothing.
2. List the jobs you want more of, honestly weighted by margin. A profitable job you would take five more of goes near the top. A headache you accept to keep the crew busy goes at the bottom.
3. Multiply service by place. Pair each service with each suburb or town you genuinely serve. Do not invent coverage - a page about a suburb you never visit also fails to rank.
4. Add the problem phrasing. Customers often search the symptom rather than the service: fuse box keeps tripping, tap dripping behind wall, jaw clicking when I chew.
5. Apply the paid ads test. Would you spend your own money on a click for this phrase? If you would not pay two dollars for that visitor, do not build your strategy around them.
6. Rank by value, not volume, then map each term to the page that should win it. Where no page exists, that gap is your content plan - a better one than four blog posts a month chosen by nobody in particular.

What a good keyword set looks like on a report

• Brand and non-brand are reported separately, and no average position figure mixes the two.
• Most terms are service plus location, in the words a customer would say out loud.
• Every term is mapped to one page, so a ranking failure points at something you can go and look at.
• The list is stable month to month, with additions explained. A list that quietly changes shape cannot be compared with last month's, which makes the report unmeasurable.
• Terms are weighted by what they are worth to the business, not sorted by search volume.
• There are some hard ones on there. A list where everything is already green was chosen to be easy.
• Somebody can tell you, without preparation, why each of the top ten terms is on the list.

That last one is the test I lean on hardest. If a list was built properly, explaining it takes two minutes.

What to do if yours is wrong

Do not fire anybody on the strength of a ten-minute check. A wrong keyword list is fixable, and how the conversation goes will tell you more than the list did. Send your five phrases across and ask three questions. Which are being targeted, and which page should rank for each? If they are not, what was the reasoning? And what would it take to get them into the plan and the report next month?

A good provider will give you a real answer, and sometimes that answer is that you are wrong. There are legitimate reasons to leave a phrase off: it might be dominated by results you cannot enter, or better served by paid ads while the organic work builds. The point is that a reason exists and gets explained.

The bad answer sounds like reassurance. It talks about long-term strategy, mentions rising impressions and domain authority, and never names one of your five phrases or one of your pages. If you get that answer twice, you are being handled rather than managed - and that is the point to read how to choose an SEO provider and the one question to ask before you start SEO before signing anything else.

One caution about the fix. Changing the keyword list means changing what pages target, and done carelessly that undoes rankings you already hold. Retarget one page at a time, with somebody watching. This is one of the ways businesses end up having paid for SEO and gone backwards.

Keyword selection is the cheapest decision in SEO and the one with the longest shadow. It costs an afternoon, and it determines what every dollar spent after it is pointed at. Get it right and ordinary work produces enquiries. Get it wrong and excellent work produces a graph. So do the ten-minute check. If three of your five are missing, you have found the problem, and you found it without understanding a single technical thing.