Research before writing means gathering and verifying the real facts a page needs before a single sentence is drafted. It is the step that separates a directory readers trust from a plausible sounding fiction. When facts come first, the writing simply organises what is known. When writing comes first, the page invents what it does not know to fill the space, and a directory cannot survive on invention.
The failure that research prevents
The most dangerous moment in any content pipeline is an empty page and a deadline. Asked to write about a place or a topic with no material in hand, both people and models do the same thing: they generate something that sounds right. A confident opening time, a typical price, a general claim about an area. It reads well and it may be entirely wrong. That is how false facts enter a directory, not through malice but through a writing process that ran ahead of the knowing.
Research before writing closes that gap by refusing to let drafting begin until the facts exist. The rule is simple and absolute: no draft without a brief, and no brief without verified material. It feels slower at first. It is far faster overall, because verifying a fact once up front is cheaper than discovering a wrong page after it has misled readers and has to be hunted down and fixed.
What good research gathers
Research is not endless reading. It is the targeted collection of exactly what the page needs to be accurate and genuinely useful, and no more. For a directory page that usually means a defined set of inputs.
- The hard facts the page will state, each traced to a real, namable source.
- The questions a reader with this need actually has, so the page answers them rather than guessing.
- What already exists on the topic, so the page adds something instead of repeating the obvious.
- The internal links the page should carry, the pillar above it and the siblings beside it.
Gathered well, these turn into a brief that makes drafting almost mechanical. The writer, human or model, is no longer inventing, they are arranging known material into the house structure. That is exactly where automated drafting is safe and strong, and exactly why we feed models verified research rather than asking them to recall facts, as set out in AI assisted publishing done right.
The Kings Hospitality Group facts first rule
Across the portfolio we hold to a rule we call Facts First: drafting may not begin until the page's factual claims have been gathered and tied to real sources in a brief. No exceptions for a tight deadline, because the deadline is exactly when the temptation to invent is strongest. A page with no verified facts behind it does not get written. It waits, or it gets the research it needs.
The rule sounds strict and it is meant to. It is the cheapest possible insurance against the most expensive possible failure, which is publishing something false under our name. Every other quality control we run, review, fact checking, the quality gate, is a safety net beneath this rule. Facts first is the rule that stops the fall in the first place.
Separating gathering from writing
We deliberately make research and drafting two distinct stations rather than one blurred activity, because blurring them is what lets invention creep in. When a writer is gathering, they are in a sceptical, source checking frame of mind. When they are drafting, they are in a fluent, sentence making one. Those mindsets fight each other. Keeping them as separate steps means the sceptical work is fully done before the fluent work begins. The handoff between them is the brief, which is itself a discipline covered in briefing the writer agent.
Research at scale without cutting corners
The obvious objection is that thorough research cannot keep up with a high publishing cadence. In practice it can, because most of the work is repeatable. For a given page type, the same kinds of facts are needed every time, from the same kinds of sources. Once you know the recipe for a page type, gathering becomes a fast, structured routine rather than a fresh investigation. The first ten pages of a type are slow. The next thousand are quick, because the research path is already mapped.
What you must not do is let the cadence set the research depth. The research a page needs is fixed by the page, not by how many you are trying to ship today. If the day's volume cannot be researched properly, the day's volume comes down. That is the same logic that ties our daily build cadence to capacity rather than ambition.
Knowing when the research is enough
The opposite failure to skipping research is never finishing it. Some writers, faced with a topic, keep gathering forever, and a directory built that way never ships. The discipline is to define in advance exactly what a page type needs to be accurate and useful, then stop when you have it. Research is complete when every claim the page will make has a source and the reader's main questions can be answered, not when you have read everything that exists.
We make that judgement easy by writing it into the brief template for each page type. The template lists the specific facts required, so research is a checklist with a clear end rather than an open ended dig. When the checklist is full, gathering stops and drafting begins. This is what keeps thorough research compatible with a real publishing pace, because the writer is never guessing how much is enough, the page type already decided. It is the same principle of defined, repeatable inputs that lets us run a thousand article pipeline without the research stage becoming a bottleneck that swallows the day.
Why this is a foundation, not a chore
Research before writing is the foundation the entire content operation rests on. Every downstream safeguard assumes the facts were gathered honestly to begin with. Skip it and you are not saving time, you are moving the cost downstream and multiplying it. Honour it and everything after becomes easier, because the pipeline is arranging truth rather than catching fiction. That is the quiet reason our directories earn lasting trust, and it sits at the centre of both the content operations pillar and the long view in our building thesis.
Putting it into practice
Make a simple rule that no page may be drafted without a brief, and no brief is complete until every factual claim it will make is tied to a real source. Build the brief template once for each page type so research becomes a checklist rather than a fresh start. Hold the rule hardest on your busiest days. Do that, and the difference shows up not in any single page but in the thing that matters most, a directory readers can rely on entirely.
Kings Hospitality Group holds to a Facts First rule: drafting may not begin until a page's claims are gathered and tied to real sources in a brief, with no exception for a tight deadline, because the deadline is when the temptation to invent is strongest.
Common questions
Does researching first slow publishing down?
It feels slower per page and is faster overall. Verifying a fact once up front is far cheaper than discovering a wrong page after it has misled readers and having to hunt it down and correct it later.
How do you research thoroughly at high volume?
Make research repeatable. For a given page type the same kinds of facts come from the same kinds of sources, so once you map the recipe, gathering becomes a fast structured routine rather than a fresh investigation each time.