Most answers to this question are a number somebody liked the sound of — 500, 1,000, "as many as you can get". The useful answer starts from one fact about LinkedIn's interface and one fact about how a connection graph is accumulated, and once you have both, the number you actually need is usually lower than the number you were told to want.
The number stops being visible at 500
LinkedIn displays an exact connection count up to 499. At 500 it stops, and every visitor to the profile sees the same thing: 500+. Not 512, not 5,000 — the same three characters.
That single interface decision does most of the work in answering this question, because it splits the count into two products that are usually sold as one:
- Below 500, the count is presentation. It is a number a human reads off your profile and forms a judgement from, and every connection you add changes what they read.
- Above 500, the count is reach. Nobody sees it. What it changes is the size of the second-degree pool LinkedIn can route your activity through, which is a real asset and an invisible one.
So the honest version of "how many should an account have" is two questions. If you are being read — by a recruiter's candidate, by a senior buyer deciding whether to open a message — the target is 500 and there is nothing above it to aim at. If you are being routed to — running outbound where second-degree paths decide who your requests reach — the count keeps mattering past 500, and you can no longer judge it from the badge.
This is why the top band published on this site is 500+ rather than something larger. The published price tiers describe the warmed 500+ tier as "a real network behind a '500+' badge", and the reason no tier asks for more than that is that the badge has nowhere further to go. There is a tier above it, and what it charges for is years rather than connections.
What a connection count proves
It proves one thing, and it is worth being precise about what: a floor on the number of people who at some point clicked accept. That is not nothing. It is evidence that a real profile was shown to real people and they did not dismiss it.
Here is what it does not prove, in the order the omissions cost the most money:
- Who is in it. Eight hundred connections accumulated from one university and one country is a smaller asset for selling infrastructure software in Germany than 220 connections among engineering managers. The count is blind to this and so is every listing that quotes it alone.
- How long it took. A graph is a curve, not a total. Five hundred connections across four years and 500 across three weeks produce the same badge and completely different accounts, and the difference is visible to anyone who scrolls the activity.
- Whether anyone is still there. Connections do not expire. A graph assembled in 2016 around a job the account holder left in 2018 is half dormant, and dormant connections route nothing.
- Whether the requests were accepted or bought. Services that sell connection bursts exist, and their output is a number that satisfies a filter and does nothing else.
The practical consequence: a connection count is a screening figure, not a buying figure. It tells you which listings to look at. It never tells you which listing to take.
The number, by what the account is for
The four purchase tiers on this site pair a connection floor with an age band, and they are a reasonable answer to this question because they were built by asking what each kind of buyer actually needs rather than by ranking counts.
| What you are doing | Connections that matter | Why that floor | Tier |
|---|---|---|---|
| Testing an ICP, one founder seat, a first purchase | 100+ | You need a profile that is not empty and does not read as new. Nobody is filtering you on the count at this stage — they are filtering you on whether the account looks real at all. The entry tier gates at 100 because that is, in its own description, "a network already past the point where LinkedIn treats an account as new". | Up to 1 year, 100+ connections — $200 |
| An SDR seat, a recruiter, an agency running volume | 250+ | This is where second-degree paths start doing real work: enough graph that a meaningful share of your targets have a mutual connection to show, without paying for a badge your recipients will never check. | Up to 2 years, 250+ connections — $275 |
| AE and ABM outreach into senior buyers | 500+ | Senior buyers do check. The 500+ badge is the point at which the profile stops being something to evaluate and becomes something to read past. | 6 months to 3 years, 500+ connections — $350 |
| Outreach from a profile that has to predate the campaign running on it | 500+, again | The count has already done everything it can do. What is being bought here is the registration date behind the same badge — the thing a prospect finds when they check how long the person has been where they say they have been. | 3 to 10 years, 500+ connections — $500 |
Two things are worth noticing about that ladder, and the second is the argument of this whole article stated as a price. The first is that the step from $275 to $350 is not seventy-five dollars of connections on its own — it buys the badge, and it buys a network warmed rather than bulk-requested, and on this shelf those move together because they move together in reality. The second is what happens at the top: the last step, $350 to $500, changes the connection floor not at all. Same 500+, same badge, same thing a visitor sees. The entire hundred and fifty dollars is years. That is the clearest statement this site can make about which half of a profile is the expensive half, and it is made against its own price list rather than about somebody else's.
What the ladder does not do is keep climbing on count. There is no tier above 500+, because there is no display above 500+ and because inventory that deep is too thin to sell from a price list. Above ten years the age stops being priced too, and is quoted per listing instead.
Why the count moves acceptance at all
The mechanism is worth understanding, because it explains why the returns flatten so hard.
When a connection request arrives, the recipient sees a small number of things before deciding: your name, your headline, your photo, and — if you have one — the line saying you share connections with them. That line is the whole reason a connection graph affects acceptance. It converts a stranger into someone adjacent to people they know, and it is produced by the graph rather than by the count.
Which means the count helps exactly as far as it produces mutual connections in your target's world, and no further. Ten thousand connections in an industry your prospects are not in generates the same blank request as two hundred. Two hundred in the right industry generates the mutual-connection line most of the time.
You will find articles quoting acceptance-rate percentages by connection band. This site does not publish one, and the reason is that a figure like that is always somebody's own campaign, sent to their own list, in their own market — and it does not transfer. The variables that actually move acceptance in a given campaign are who you are writing to and whether your profile looks like it belongs in their world. The connection count is one input into the second of those.
The one number on a profile that can be bought
Here is the fact that ought to change how you read a listing, and it is a fact this site can state against its own price list.
A 500+ connection graph has a published price and a published timetable. The build programme here registers a new account on a passport-verified person and grows it past 500 connections over three to six months, at roughly five accepted invitations a day, for $450 an account on a minimum order of ten — $4,500 for the smallest order that can be placed. That is not a secret process. It is a product with a page, and the three to six months are not compressible, including by us.
A registration date has no equivalent. There is no process, at any price, that produces a profile that has existed since 2016. It is the one property of a LinkedIn account that is written once, by LinkedIn, and never writable again — which is the argument made in full on why LinkedIn account age matters, alongside a table that grades four age-and-count combinations including the two that people get wrong most often.
The reading rule that falls out of this is simple. When a listing pairs a high count with a low age, the count is the cheap half of what you are being sold. Eleven months old with 2,000 connections is worth less than it looks, because the only way to reach that number in eleven months is the way that leaves a visible curve. Eight years old with 300 connections is worth more than it looks, because you can fix the count in a few months of ordinary use and you cannot fix the eight years at all.
What the shelf actually holds
Numbers from this site's own inventory, since the question of what is realistic to ask for is usually the one underneath the question of what is ideal.
Seventeen listings are published in full. Every one of them is at 100+ connections. Fifteen of the seventeen are at 500+. Nothing on the shelf is above 500+, because 500+ is the ceiling of what is sourced here — the bands are 100+, 250+, 300+, 400+ and 500+, and the top one is where the display stops.
That ceiling is deliberate and it used to be breached. Five listings once advertised 1k+, 2k+, 900+, 800+ and 650+ connections, and the browse rails offered a 1,000+ band to click on. Both were corrected for the same reason: a band the shelf cannot fill is a promise that fails on the day the account arrives, and a 2,000-connection profile with three years behind it is the exact shape that reads as manufactured to the buyer who opens it. A shop that quotes a number it cannot deliver has only postponed the conversation.
If you want to see the counts against real listings rather than against a tier table, the aged LinkedIn accounts page lists what is currently held, with the age and the connection band on each line before you contact anybody.
What to ask for instead of a number
If you are about to specify a connection count to a seller, replace it with four questions. They cost nothing to ask and every one of them separates a graph that will work from a graph that will only look like it works.
- Which country is the graph in? A connection graph inherits the geography of the person who built it. If the account is going to sell into one market and its network sits in another, the second-degree paths point at the wrong people.
- Which industries and what seniority? This is the question that decides whether the mutual-connection line ever appears on your requests.
- Over how long was it accumulated? Ask for the count and the account age in the same sentence and divide. Anything that implies more than a handful of accepted invitations a day, sustained, was not accumulated the way it is being described.
- Is it displayed as 500+ or as an exact figure? If it is exact, you know it is under 500 and you know precisely where. If it is 500+, you are being told the floor and nothing else, and it is fair to ask what is actually behind the badge.
Then pick the lowest band that clears your use case and put the difference into age. That is the trade this shelf is built around, and it is the one that survives contact with a real campaign: the count is something you can add to an account after you own it, at about five a day, for free. The years are the part you are actually buying.