Negative keywords across ten boroughs

A Google Ads account that covers Greater Manchester needs two negative keyword lists that a small-town account never thinks about: the place names inside the county that the business does not serve, and the words that mark a search as coming from one of the city's students or its job seekers rather than a customer. A plumber in Stockport does not want 'plumber wigan'; a Deansgate law firm does not want 'law degree manchester'. Both searches match the keywords, both cost money, and both are stopped by a negative list built for this county rather than copied from a template.

The place-name problem

Location targeting decides where the searcher is, not what they typed. Someone standing in Sale who searches 'emergency electrician bolton' is inside your Trafford radius and outside your service area at the same time, and Google will show your ad. In a county with ten boroughs, several large towns inside each, and a city-centre postcode that half the region uses as shorthand, the searches that name the wrong place are a steady leak. The answer is a shared list of the places you do not go, applied to every Search campaign.

Build it from the map, not from the Search Terms report alone. List the ten boroughs and the towns inside them, mark the ones you serve, and put the rest in the list at phrase match, so 'plumber wigan' is blocked but a search for a road that happens to contain the word is not. Keep the list in a spreadsheet as well as in the account, with the reason beside each town, so the next person to open the account knows why Prestwich is in and Whitefield is out. Then add the neighbours that the radius will reach anyway: Warrington, Macclesfield, Glossop, Huddersfield, Blackburn, Chorley, and Liverpool for anyone west of Salford.

The third column is one business's answer, not a rule. A Bolton roofer reverses most of it. The point of the table is that the decision is made town by town, in writing, before the first click, and revisited when the Search Terms report shows a converting enquiry from a town on the negative list.

The student and job-seeker problem

Manchester is a university city several times over, with the University of Manchester, Manchester Metropolitan, the University of Salford, the Royal Northern College of Music and the University of Bolton all inside the county, and every one of their students searches for jobs, courses, accommodation and cheap versions of things. A search for 'graphic design manchester' is as likely to come from a second-year student as from a business that needs a logo. The negative list for a city account therefore starts with education and employment words at broad match: course, courses, degree, university, uni, lecture, student, placement, internship, apprenticeship, job, jobs, vacancy, vacancies, salary, hiring, recruitment, cv.

Two more themes follow from the same source: accommodation words for any business whose keywords touch property, and 'cheap', 'free', 'discount', 'student discount' for retail and hospitality, where a student search may still be a customer but a much smaller one. The second group goes in at phrase match and is reviewed against conversions, because 'cheap' is sometimes exactly the customer a business wants.

Where the lists live and how they match

We run negatives as shared lists at account level, one per theme: places we do not serve, education and jobs, information searches, products we do not sell, competitors we choose not to bid against. Each is applied to every Search campaign, and a new campaign is not switched on until the lists are attached. Ad group negatives are reserved for shaping which ad group a search reaches, not for excluding it from the account.

Negative match types behave more strictly than positive ones. A negative at broad match blocks any search containing all its words in any order; at phrase, the words in that order; at exact, only that search on its own. Negatives do not stretch to close variants, so plurals and misspellings need adding separately: 'job' does not block 'jobs', and 'stockport' does not block 'stockport road' unless you want it to.

The weekly read, for a county-sized account

1. Open the Search Terms report for the last seven days, sorted by cost, and read every row that spent without converting.
2. Sort the wasted rows into the five lists. A place name goes to places; 'course' goes to education; 'what is' goes to information.
3. For each place name, check whether it is inside or outside the radius. Inside means the negative alone will not stop it, so the location settings need a look too.
4. Add each term at the widest match type that will not block a wanted search. Place names go in at phrase.
5. Read the converting rows last, and look for a town on the negative list. One conversion from Ramsbottom is a data point; three is a reason to serve Ramsbottom.

A place-name list is never finished, because Greater Manchester's towns have districts of their own and searchers use all of them. What matters is that every week the list gets a little longer and the account pays for a little less of the county it does not serve.

Our free generator produces the education, jobs and information lists for a named sector, and if you want the place-name list built for your boroughs as well, send us what you do and where you go.