Keyword research for Google Ads is where most small business accounts either go right or waste money fast. The trick is not finding the most phrases. It is deciding which handful of searches represent a person who is ready to buy from you, and being willing to ignore everything else. The process below works for a single-city service business or for a business selling across India.
Start from what customers actually call you
Write down the words your team uses on the phone. Customers describe problems, not services, so the phrase that produces enquiries is often closer to a symptom than to your product name. A plumbing company may sell "pipe repair" but be found for "water leak inside wall". A B2B consultant may sell "revenue operations" but be searched for as "sales pipeline not working". Ask five recent customers how they would describe the problem in their own words, then compare that with what you would have typed.
Keep a shared sheet with three columns: the phrase, the service it maps to, and whether you can serve that customer profitably. The third column does the real work, because a phrase that maps to a service you no longer offer, or a location you cannot reach, is wasted spend no matter how popular it is.
Use the tools to generate, not to decide
The keyword planner inside a Google Ads account, autocomplete suggestions and related searches are useful for generating candidates. They are also biased towards volume, which is not the same as value. A phrase with modest volume and clear intent can produce better enquiries than a high-volume phrase that attracts students, competitors or people who only want information. Treat every tool output as a list of questions to test.
Look at what competitors appear for. If a large brand occupies every position for a broad commercial term, that term may be impractical for a small budget. That does not mean avoid it, but it does mean expect to bid differently and accept a longer test period.
Check intent, then group by theme
Sort candidates into three buckets: ready to buy, researching, and not relevant. Ready-to-buy phrases often contain a service word, a location, a price signal or a direct need. Researching phrases are useful later, usually for content rather than for a first campaign. Irrelevant phrases belong on the negative list from day one.
Group what remains by theme, not by keyword tool. One ad group should be something you could describe in one sentence, with ad text that speaks to that exact situation. This is also the structure that makes reporting readable later.
Build a negative list in parallel
The fastest way to waste an ad budget is to pay for searches you never wanted. Add negatives for unrelated terms, free, cheap, jobs, training, DIY, used, wholesale when you are not wholesale, and competitor brand names if you have decided not to bid on them. Add negatives for your own services that you cannot serve in the chosen area. Review the search terms report weekly at first, because a budget without negative keyword maintenance will drift towards wasted spend within weeks. Our guide on negative keywords goes deeper into this.
Test in small, comparable groups
Create separate ad groups for themes you genuinely believe differ, such as one service versus another, or one city versus another. Keep the bids, landing pages and ad style as similar as you can, so that when one group outperforms the other you have a reasonable hypothesis. Change one variable per test. If you change the headline, the landing page and the bid in the same week, the result will not teach you anything.
Give each group a fair run before you judge it. A small daily budget may not generate enough data in a week, and a group that looks weak on day two may simply have been unlucky. Decide the test length in advance and write it down.
Questions to ask whoever does the research
Ask which tool produced the keyword list and whether the tool data was used as a filter or a target. Ask which keywords are excluded and why. Ask what the expected cost per click is for the top five phrases and what that implies for your monthly budget. Ask which phrase they expect to convert best and what the plan is if it does not. A provider who can explain the reasoning in plain language is more useful than one who hands you a spreadsheet with no explanation.
Tirath Digital Solutions handles this process as part of Google Ads management, where management starts at INR 12,000 per month with ad spend and any third-party platform subscriptions billed separately. If you would rather do the thinking yourself, start with the Google Ads guides and bring questions through our contact page.
Common questions
How many keywords do I actually need for a small account?
Enough that you can group them clearly and afford to bid on them. Many small accounts run well with a few dozen carefully chosen phrases across two or three themes, because the budget can be spent properly on each. A large list is only useful if you can afford to give every group a fair test. Extra terms that nobody searches for add noise rather than reach.
Should I use broad match or exact match keywords?
Both have a place, and the honest answer is to test rather than argue. Broad match can surface useful long-tail searches and needs tighter negative keyword control, while exact and phrase match give you clearer control over cost. Start with match types you can supervise, add broad terms in a separate experiment, and judge them on enquiries rather than on search terms volume.
Where do I find real customer wording?
Ask your sales team, read the questions customers ask on WhatsApp, check your call notes, and look at the search terms report inside your own account. Auto-suggest and related searches help, but they are suggestions, not evidence of buying intent. Treat every list you collect as a hypothesis, then confirm it against what actually converts inside the account.