How to pick keywords with no ranking data to start from
A method for choosing search terms when your site has no history, using evidence you already own instead of volume numbers you cannot verify.
Starting from nothing
A new site has no Search Console data because nothing ranks yet. That is the ordinary starting position and it is the one every keyword tool handles worst. The tools will happily hand you a list of several thousand terms with volumes attached and none of those volumes were counted. They are modelled estimates and for Indian long tail phrases they are frequently rounded to zero even where real demand exists.
So the task is not to find the terms with the biggest numbers. The task is to work out which searches your buyers actually perform and which of those a page you can honestly write would satisfy.
What research cannot tell you and what it costs
Keyword research cannot tell you whether you will rank. It describes demand and says nothing about your ability to compete for it. Difficulty scores are a vendor's model rather than a measurement. Treat them as a rough ordering and never as a forecast.
The free sources cost only time and they are the better sources at this stage. Paid tools mostly bill in dollars so the rupee cost moves with the exchange rate and a single month of access is usually enough for a first list. The larger cost is the publishing work that follows. A research exercise producing five hundred terms is worthless to a business that can publish two pages a month. Buy the list that matches your capacity to act on it.
Who needs this and who does not
This matters for any business whose buyers describe their problem in words before they buy. Services products training equipment and anything researched rather than grabbed.
It does not matter for a business selling to a known list of fifty buyers through relationships and tenders. The search demand there may genuinely not exist and a month of research will reach that conclusion expensively. A local trade with one shop does not need a keyword set either. Three clear pages and a well maintained map listing cover almost all of that demand and a two hundred term plan will sit unused. If your constraint is that nobody can publish anything then keyword research is not the purchase to make this quarter.
Where the evidence actually comes from
Start with records you already hold. The questions people ask on sales calls. The wording of enquiries that arrive on WhatsApp and email. Your own site search log if the site has one. These are real buyers using real language and no tool has better data about your market.
Then widen it cheaply. Type the core phrase into Google and read the autocomplete suggestions and the related searches at the foot of the page. Both are drawn from genuine query behaviour. Look at the page titles rivals have built because a page is evidence somebody judged the demand worth serving. Use Keyword Planner for broad ranges while remembering that it reports ranges rather than counts.
Now sort the list by intent. Separate people ready to buy from people comparing from people with a problem they have not named yet. Check who currently occupies the first page for each term. If it is filled by marketplaces aggregators and national brands then treat that term as a long project rather than a first move. Pick a small set. A few terms that match what you sell. A larger set of specific questions you can answer properly. One or two you expect to lose so you learn something.
How you know the choices were right
The first honest signal is impressions appearing at all for the terms you chose. That means Google has understood what the page is about even if it is placing you on page five. The signal that matters after that is non brand clicks and the enquiries that follow them.
The flattering numbers are the total search volume of your list the count of terms ranking somewhere in the top hundred and the difficulty scores you cleared. All three can improve while enquiries stay flat. Within a few months real Search Console data should be replacing your estimates entirely. A plan still being justified by tool volumes after six months of live data is not being read.
The mistake that costs months
The common error is choosing by volume. Head terms carry the biggest estimates and they are contested by sites with years of accumulated history and budget to match. Months of writing go into pages that never surface while the specific phrases that describe exactly what you sell go unwritten. Those specific phrases convert better anyway because the person using them has already decided what they want.
The second version is picking the wrong language. Many Indian buyers search in a mix of English and transliterated Hindi or a regional language and the volume tools under report it badly. Targeting a national term when you serve one city has the same effect. Nobody can promise a ranking for any term you choose and nobody can tell you in advance which of these choices will work. The list is a hypothesis and live data is the only thing that settles it.
This guide explains how something works. It is not legal, tax or financial advice, and rules change. Confirm anything regulatory with a qualified professional before you act on it.