Most business directories ask you to trust them. They tell you they are well organised, that their listings are good, that being listed with them is worth your time. Almost none of them show you the evidence.<\/p>\n
We decided to do the opposite. Over the past months we took our entire database — every one of the 14,362 listings in the Jasmine Business Directory — and analysed it properly, the way a researcher would analyse any large dataset. Then we published the results in full: five detailed studies, with the methods written out so that anyone can check them.<\/p>\n
This article is the short, plain-language guide to what those studies found. It is written for people who are not going to read five long research papers, and who should not have to. If you have a listing with us, it tells you what our data says about listings like yours. If you are deciding where to list your business<\/a>, it shows you how to read a directory honestly — ours included. And if you simply want to understand how a directory is actually built, over years, by human decisions, the studies turned out to be a clear window into that, and this article opens the window for you.<\/p>\n Nothing here asks you to take our word for anything. Every number in this guide comes from a published study, and every study shows its work. That is the point. A directory that has measured itself and put the results in the open is making a different kind of claim than one that simply describes itself in comfortable language, and the difference is worth a few minutes of your reading.<\/p>\n Jasmine Business Directory has run since 2009. Two things about how it works shaped what we wanted to know, and they are worth stating before any of the findings, because they explain why our database is worth reading at all.<\/p>\n The first is that we have never used paid advertising to bring in listings. The directory did not grow because we bought traffic or paid for placements. It grew because businesses found us and submitted themselves, or because our editors found businesses worth listing and added them. Every entry in the database arrived through one of those two routes, and that makes the database an honest record rather than a bought one.<\/p>\n The second is that roughly nine in ten of our listings were added by hand, through editorial review, rather than through an automated feed. A human looked at each of them. That matters because it means our database is not a machine-generated scrape of the web; it is the accumulated result of seventeen years of editorial decisions about what belonged in the directory and what did not. A record like that has a history<\/a> written into it, and a history can be read.<\/p>\n There is one more reason a directory like ours is worth analysing rather than just describing. Because nothing was bought and most of it was added by hand, the database is unusually close to a record of genuine demand and genuine editorial judgement. It is not padded with automated entries that exist only to inflate a number. Whatever the studies found — flattering or not — they found it in a corpus that was built the slow way, and that makes the findings worth taking seriously rather than dismissing as an artefact of how the listings were gathered.<\/p>\n We also wanted an honest picture rather than a flattering one. It is easy to describe a directory in round, comfortable numbers.<\/p>\n It is harder, and more useful, to ask what the data actually shows — including the parts that are not flattering. The five studies were written in that spirit. They are descriptive: they report what our corpus is, in detail, without dressing it up. Some of what they found is genuinely good, and some of it gave us a clear and uncomfortable list of things to fix. Both are in there, side by side, because a picture with the awkward parts edited out is not worth the effort of making.<\/p>\n You do not need to read all five studies to get the value from them. That is what this article is for. But they are all published, the full versions, if you want to go deeper — and a directory that publishes its own weak points is, we would argue, a directory worth a closer look.<\/p>\n Before the findings, it is worth a moment on how the studies were done, because the method is what separates a real analysis from a marketing<\/a> claim dressed up with numbers.<\/p>\n The studies looked at the whole database, not a sample. There are 14,362 listings in the directory, and the analysis covered all 14,362. This sounds like a small detail, but it removes a whole category of doubt.<\/p>\n When a study works from a sample, you have to ask whether the sample was representative, whether it was chosen to flatter, whether a different sample would tell a different story. When a study works from the complete set, those questions disappear. The numbers are not estimates of what the directory is like; they are what the directory is.<\/p>\n The studies are also descriptive rather than predictive. They do not claim to forecast anything or to prove a theory. They report measured facts: how many listings sit in each category, how many fields each listing has filled in, when each listing was added, where each listed business is. A descriptive study has a narrow job, and the narrowness is a strength — it means the findings are about as solid as findings get, because they are counts of things that are either true or not.<\/p>\n And the methods are written out in full. Each study explains exactly how its numbers were produced, step by step, in enough detail that someone with the same database could repeat the work and get the same results. That is the ordinary standard for research, and it is an uncommon one for directories<\/a>. It is also the reason you do not have to trust us: the studies are built so that they can be checked rather than believed.<\/p>\n None of this makes the studies difficult to read in summary, which is the whole premise of this article. But it does mean that the short answers below rest on something firmer than a confident tone. They rest on a complete count of a real database, done in the open.<\/p>\n The series has five parts. Four of them each examine one property of the directory; the fifth pulls everything together. Here is what each one asked and what it found, with the technical detail left out.<\/p>\n The first study looked at how our 14,362 listings are distributed across the directory’s categories. The short answer is: very unevenly. A small number of categories — Law<\/a> above all, but also areas like cosmetic procedures, home improvement, and business and finance — hold a large share of all listings. Hundreds of other categories hold only a handful each. The hundred largest categories, out of several hundred in use, account for just over half of everything in the directory.<\/p>\n This unevenness is not unusual. Almost any large collection organised by category ends up lopsided, because the world itself is lopsided — there are simply more law firms and more home-improvement businesses seeking visibility than there are, say, specialist antiquarian map dealers. The study measured exactly how lopsided our directory is, and put a single clear number on it, so that the shape is documented rather than guessed.<\/p>\n This matters to you because it tells you how much company your listing keeps. In a crowded category, your listing sits among many competitors and has to work harder to stand out. In a quiet category, the same listing is far more visible simply because there is less around it. Neither is better in the abstract; the point is to know which situation you are in, because it changes what your listing has to do to be noticed.<\/p>\n The second study asked a simple question: of the core pieces of information a listing can carry — its name, description, keywords, address, contact details, and so on — how many does a typical listing actually have filled in? We counted ten such fields for every listing.<\/p>\n The average came out at 3.67 fields of ten. But the average turned out to be misleading, and that misleading average became one of the most important threads in the whole series.<\/p>\n Listings did not cluster around three or four filled fields. Instead they piled up at the two extremes: a very large group of listings with almost nothing filled in, and another large group with almost everything filled in. Few listings sat in the middle. That split — near-empty on one side, near-full on the other, and a hollow space between — is the heart of the big finding, and the section below returns to it. For now the point is just that an average can describe a collection accurately and still describe almost none of the things in it.<\/p>\n The third study looked at geography — where in the world the listed businesses actually are. Only about a third of our listings carry a country at all. Among the ones that do, the spread is heavily concentrated: the United States and the United Kingdom<\/a> together account for the large majority, and the English-speaking world as a whole accounts for nearly nine in ten located listings.<\/p>\n So the directory has a clear geographic shape. It is strongest in the United States<\/a> and the United Kingdom, useful across the rest of the English-speaking world, and thin elsewhere. Within the United States, the listings lean toward a handful of states — California and Florida<\/a> most of all — and toward large cities, with London the single most common city across the whole directory.<\/p>\n If your business and your customers are in those strong regions, the directory’s geography works in your favour: you are listed where the directory is densest and most used. If you are outside them, it is still worth knowing the shape, because it tells you what kind of company your listing keeps and how the directory is most likely to be searched.<\/p>\n The fourth study reconstructed when every listing was added, year by year, from 2009 to today. Growth was not steady. The first three years were quiet. Then 2013 and 2014 were enormous: those two years alone account for almost exactly half of every listing in the directory. After 2014 the pace dropped sharply and has stayed at a few hundred listings a year since.<\/p>\n That 2013–2014 surge was a deliberate, intensive period of editorial work, and it lines up exactly with the years the directory won a series of awards for its editorial curation. It is a real and documented chapter in the directory’s history, not an accident in the data.<\/p>\n It also means most of our listings are now more than a decade old. The typical listing in the directory was created during that intensive period, which has consequences — for how current the information in it is likely to be, and for the big finding — that the sections below come back to.<\/p>\n The fifth study is different from the other four. The first four each measured one property on its own. The fifth crossed all four against each other, listing by listing, to ask whether they were related — and found that they were, so closely that they turned out to be one fact seen four ways. That fact is the subject of the next section, and it is the single most important thing the series found.<\/p>\n The fifth study did something the first four could not. Each of the first four looked at one property on its own — categories, completeness, geography, time. The fifth study crossed them against each other, listing by listing, to see whether they were related. They were. In fact they turned out to be the same fact wearing four different faces.<\/p>\n Here is the fact. Our directory is really two directories sharing one database.<\/p>\n One of them is the older directory: the listings added between 2009 and 2014, which make up about 63% of everything we hold. These listings are mostly skeletons. On average they have fewer than two of the ten core fields filled in. Most carry no location at all. They are little more than a name and a category — a placeholder where a full listing should be.<\/p>\n The other is the newer directory: the listings added from 2015 onward, about 37% of the total. These are a different animal. On average they have nearly seven of the ten fields filled in, and seven in ten of them carry a proper location. They are, for the most part, real, complete, usable listings.<\/p>\nWhy we analysed our own directory<\/h2>\n
What makes this kind of analysis worth trusting<\/h2>\n
The five studies, in plain terms<\/h2>\n
Study one: how listings spread across categories<\/h3>\n
Study two: how complete the average listing is<\/h3>\n
Study three: where the listed businesses are<\/h3>\n
Study four: how the directory grew over seventeen years<\/h3>\n
Study five: how the four findings fit together<\/h3>\n
\n
\n \nStudy<\/th>\n The question it asked<\/th>\n The short answer<\/th>\n<\/tr>\n<\/thead>\n \n One<\/td>\n How do listings spread across categories?<\/td>\n Very unevenly — a few categories hold most listings<\/td>\n<\/tr>\n \n Two<\/td>\n How complete is the average listing?<\/td>\n Most are either nearly empty or nearly full, with little in between<\/td>\n<\/tr>\n \n Three<\/td>\n Where are the listed businesses?<\/td>\n Mostly the US and UK; nearly nine in ten located listings are English-speaking<\/td>\n<\/tr>\n \n Four<\/td>\n When were the listings added?<\/td>\n Half the directory was built in 2013–2014; most listings are over a decade old<\/td>\n<\/tr>\n \n Five<\/td>\n How do these four findings fit together?<\/td>\n They are all one underlying fact: the directory has two layers<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n The big finding: the directory is two directories<\/h2>\n