Building a collection that makes sense

Start with the uncomfortable part: most people do not need more than one watch, and a good one will do everything they ask of it for decades. A collection is not a solution to a problem. It is a hobby, and it is worth being honest that you are choosing to spend money on a hobby rather than solving anything.

That honesty makes the rest easier. Once you stop pretending the second watch is a necessity, you can ask a better question: what would make a group of watches genuinely interesting to own, rather than just more of them? The answers that work tend to be about coverage or about a theme, and almost never about climbing a ladder of brands.

Coverage: the sensible default

The classic answer to "which three watches?" is a coverage answer. One watch you wear every day, one that survives water and knocks, and one that works with a suit or an evening. Between them they cover almost every situation an ordinary life produces, and they do it with three watches rather than nine.

Coverage is a better organizing idea than brand-chasing for a simple reason: it has a stopping point. "One of each thing I actually do" terminates. "The next brand up" does not, because there is always a next brand.

Organizing ideas that produce a more interesting collection

Coverage is sensible. Themes are more interesting, and they cost less to pursue seriously.

A theme is any constraint you impose on yourself: one decade, one complication, one country, one movement family, one case shape, watches made in the year each of your children was born. The constraint does the work. It gives you a reason to say no, it makes each purchase relate to the others, and it turns shopping into research.

The most rewarding version of this is collecting for the mechanism rather than the badge. If you buy three watches built around the same well-documented movement, you learn exactly how much of a watch's character comes from the case, the dial and the finishing rather than from what is ticking underneath. Our guide to how a mechanical watch works is a decent map of what to look at, and small brands are unusually good hunting ground for this, as our guide to microbrands and independents explains.

One in, one out

The discipline that saves more collections than any other is simple: nothing new arrives until something leaves.

It works for three reasons. It forces you to rank what you own, which is uncomfortable and clarifying. It caps the total, which caps servicing, insurance and the slow guilt of owning things you never touch. And it makes each purchase a real decision rather than an addition to a pile, because you have to name the watch it displaces.

The rule does not have to be absolute. Plenty of people run it with a fixed number of slots instead: six watches, and a seventh means choosing which six you keep.

Cadence, and the waiting period

Buy slowly. Almost nobody regrets a watch they waited three months for.

Impose a waiting period between wanting and buying. Thirty days is enough to kill most impulses. During those thirty days, do the boring work: find the reference number, read what owners say about it after two years rather than after two days, check what a service costs and who can do it, look up what the same watch has actually sold for second-hand, and try it on if you possibly can.

A useful side effect is that a waiting period breaks the link between a watch and a mood. A great many watch purchases are really a response to a bad week, and those are the ones that get sold at a loss six months later. The arithmetic of that loss is set out in our guide to watch value and depreciation.

Two traps worth naming

The consensus watch. Every few years the hobby converges on a small set of watches that everyone agrees you should own. Sometimes they are genuinely excellent. But consensus is also how you end up with a collection that looks like everybody else's, bought at the moment the price contained the most enthusiasm. If your reason for wanting something is that everyone says it is the obvious choice, that is a reason to wait, not a reason to buy.

The audience that is not looking. People buy watches to signal things to colleagues, strangers and relatives, and those people are, with rare exceptions, not looking at your wrist and could not identify what is on it if they were. A watch is a thing you look at, forty times a day, at a distance of half a meter. Buy for that person. Purchases made for anyone else have the highest regret rate in the hobby.

Trying before you buy when you cannot see one

Most of the interesting watches in the world are not in a shop near you. Some workarounds are better than others.

• Measure the watch you own that fits best, then compare case diameter, thickness and above all lug-to-lug length. Lug-to-lug tells you far more about fit than diameter does, which is the argument our sizing and fit guide makes at length.
• Print a paper template. Several manufacturers publish scale drawings. Printed at true size and taped to your wrist, they answer the "is it too big?" question surprisingly well.
• Find a proxy. If a friend, a shop or a watch meet has anything with similar dimensions, try that instead. Dimensions transfer between brands.
• Watch video reviews for wrist shots with the reviewer's wrist size stated, and ignore the studio photography, which is designed to flatter.
• Ask about the return policy first. For an online purchase this is the real safety net, and it is worth choosing a seller partly on that basis.

Records, insurance and security

Keep an inventory. It is fifteen minutes of work per watch and it is the difference between an insurance claim that is paid and one that is argued about.

For each watch, record the brand, model, full reference number, serial number, purchase date, where you bought it, what you paid, service dates and what was done, and photographs of the watch, the caseback and the full set of box and papers. Keep it somewhere that survives a house fire or a burglary, which means a copy off the premises or in the cloud.

That inventory is close to exactly what an insurer will ask for. On insurance itself, three practical points. These are general observations rather than advice about your policy: cover varies by insurer and by country, and the policy wording is the only thing that decides what you are actually covered for.

Second, cover away from the home is often a separate option. A watch stolen from your wrist in a car park is not a contents claim in the ordinary sense, and "personal possessions" or "all risks" cover is what handles it.

Third, insurers price on replacement cost, so a valuation from a qualified valuer keeps the sum insured realistic. Revisit it every few years, because replacement costs move.

Security is unglamorous but effective: do not post photographs that show your address, your car or your routine, do not wear an obviously expensive watch in situations you would not carry the same value in cash, and keep serial numbers somewhere you can hand them to the police in ten minutes. Stolen watches are recorded in trade databases, and a serial number is what makes recovery possible at all.

Rotation, and knowing when to let go

A watch that never gets worn is not a collection, it is storage. Once a year, look at what you have actually put on your wrist. If something has not been worn in twelve months, there is usually a reason: it does not fit, it does not suit your life, or you bought it for a version of yourself that did not turn up.

Sell it. The money is better in the watch you do wear, and every unworn watch is still costing you insurance and eventual servicing. Our guide to selling a watch covers how to do that without losing more than you need to, and the servicing and care guide explains why unworn watches do not stay young in a drawer.

Leaving the hobby is allowed, too, and it is easier if you have kept records. Sell in an order that suits you, keep the one or two you would actually miss, and do not let sunk cost persuade you to keep a collection you have stopped enjoying. Plenty of people cycle out of collecting for a decade and back into it later with much better judgement.