The advice, especially to writers, is so common that it has been reduced to a meaningless cliché. Nevertheless it is, in certain cases, good advice – the key is to know when having less is more for your story and when having more is less (work) for your story. This time we’re going to focus on the first of those (I’ll explain the second another time).
1. When you use the right word, not its second cousin. The funny thing about writings is that, oddly enough, you have to love words as well as stories. All words carry connotations with them – subtle things that are used and noticed subconsciously by most people, but which can totally alter the way something is viewed. It also has a huge effect on how smooth or clunky a work is, as a writer who chooses wisely can drastically reduce the amount of words they need to say something – which in turn allows them to include far more information (depth, characterisation, plot, setting, etc) in a work of the same size. Writing is about making every word count. It might seem like a good idea to describe the universe as being a light, creamy, white-brown, but it is much easier for readers to know what you mean (and for your word count) to just to admit you’re not describing a particularly exciting colour and just use beige. If you’re willing to expand your vocabulary a little – and trust that your readers are not gibbering idiots who cannot understand if they ever have to look up a word or two, provided that its place in the text makes its approximate meaning clear – you will find that just about everything does have a word to describe it in English (and other languages). Why waste time and reader patience with a cliché like brightly coloured flowers when you can say zinnias? The key here is that using the right word rather than a less accurate word (why say someone is very irritated when they are actually wroth?) and specificity together allow an author to use very few words to create a deep and complex image for the reader, rather than something which is generic.
2. When you have people plotting. Real manipulators know better than to use many-stepped plots. (Also, caveat, not all manipulation is negative – we just tend to use other words for it or avoid admitting to what’s involved when we talk about keeping people away from surprise birthday parties, talking down the suicidal, etc.) If you’ve ever heard the quote “no plan survives contact with the enemy” you’ll know precisely why plots should be kept as simple as possible. Yes, readers might be able to follow something as complex as “To gain the throne [motive] I will: [plot] convince my lover to murder her husband so that we can be together  and to send a warning to her sister to make them suspicious of [other people] so that my lover’s family and [other party] will go to war , so that I can gain the rank needed to marry my lover  only to kill her because it was her sister who I got killed I really wanted and she’s delivered me her niece who’ll do , and who I’ll marry off to my (now late) lover-wife’s heir’s heir , and then have the heir killed so the heir’s heir and my lover-niece can inherit , and then take them to war  since the war I started exhausted everyone else’s armies, then – once that’s won – place heir’s heir and niece-lover on the throne , kill off heir’s heir once heir’s heir has an heir  and marry grieving niece-lover , then kill off heir’s heir’s heir  and inherit throne”. Each step there has, on average, twenty things that could go wrong, easily five potential witnesses who would need to be silenced, and enemies to come into contact with every step of the way because each manipulated person’s freewill and potential to do the unexpected is an extra enemy (and no true manipulator claims to be always able to know what someone will do, true manipulators just know to keep on their toes so that they can compensate when someone inevitably goes off the rails). A wiser plan than the above would be “Step one: sow political chaos in realm so advancement opportunities arise, Step Two: ???, Step Three: Profit”. This is because you really do need to re-evaluate every step of the way, so every plan should essentially be one step, because then people and coincidences and random pebbles that cripple horses (for want of a nail, etc, etc) alter everything and you need to start planning again.
The first of those plans is a theory someone gave for what Littlefinger is up to in A Song of Ice and Fire, but the second is likely a much more realistic version of his plans, because Littlefinger is canonically a good manipulator. See, here’s the thing: to an outsider, when a plot seems to have finished and reaped its rewards, it can seem like the manipulator planned it all out carefully from the start. But it just isn’t so – it looks like that because the non-manipulator only sees the final product, so if you’re going to write about a character explaining their plot it’s best that they have something very simple and then improvise and re-plot the rest of the way. Writers and tacticians often use chess metaphors to describe plotting, but here’s the thing writers tend to forget: your manipulator is not playing chess against themselves. Real chess players re-evaluate and often change strategies during play. This holds true whether you’re trying to win a political victory or ensure your prophesised child hero/sacrificial lamb permanently kills the Dark Lord you never managed to permanently kill (that Dumbledore’s plotting succeeded was sheer dumb luck; it’s an excellent example of a plot that shouldn’t have worked but somehow, due to author meddling and ignoring all the highly probable things that could have screwed it up from the beginning, it did – which is a BAD THING, by the way). It holds true regardless of whether you’re trying to talk down a suicidal person, trick a confession out of someone, take over the world, discredit an evil rival, or just about anything else where having to plot or manipulate can be involved. Oh yeah, and it also holds true if you’re planning to have a situation where something was made extremely convoluted so that people be unable to do something – for instance, the wisdom of people who lock things with fifteen special keys that all must be turned at once to open the door to the secret temple might seem like a cunning plan, but a wise manipulator would know that eventually someone’s going to come along who thinks the best solution to everything is TNT.
3. When you’re describing a person. Most people who go looking for writing advice already know better than to describe – especially as an introduction – someone’s physical appearance in lurid detail. Some people still do block-dump physical descriptions at the start of their work, for instance most of the Potter books give a block-dump description of what the titular character looks like (green eyes, glasses, messy black hair, knobby knees/small build, and plot-important scar) near the start, but they are typically wise enough to confine it to a line or two and then not really bring up appearances again for the rest of the work. This is the “give a clear indication of what they look like so readers can imagine it and then get on with the plot” method of describing characters. It’s hardly the only, or even the best, method – although it gets points for treating a physical description as an unfortunate necessity rather than an important event to be lavishly and time-consuming-ly covered from every angle.
Here’s the thing: giving a physical description of a character as something separate from describing the way they make people feel and their own nature isn’t obligatory – or even a good idea in all situations. This doesn’t mean you should fall into the good looking = good fallacy of clichés, but you can say a hell of a lot more about a person by mixing them together. Here’s the other thing: you need to choose what is important to describe – that which is interesting or of note – rather than just describing the standard description items (hair colour, eye colour, skin colour, fitness level/body shape, occasionally face general shape and nose type). I mention them one after another because it’s easier to show them as they work together. Which of these gives you more of a feeling for the character?
“She had dark brown hair, which fell to her shoulders. Her brown eyes were set in a narrow face with high cheekbones. She was tall, thin, and beautiful. She was also cruel.” (32 words)
“The dark woman would have been beautiful if only her nature were not so obvious in the coldness of her sharp features.”(22 words)
The one tells you the typical things about what the character looks like. The later tells you something about who the person is and what they look like. Things like eye colour and hair colour can easily be slipped in at other points (“she brushed a stray lock of brown hair away from her face”, for instance). When giving description it always helps to make them do double duty – to make them give both appearance and personality, or backstory (or triple duty). Everyone has hair and eye colours. Not everyone has nibbled nails or lines from frowning regularly. A few less common descriptors can give a far better impression both of the appearance and nature of a character than a lot of common descriptors.
4. When you’re giving backstory. Backstory shapes behaviour. This means that the best place to show backstory is in behaviour. For example: if you have a character whose little sister died, you don’t need them (especially if they’re taciturn by nature) to go into a long spiel about how it happened; you can imply in the way they mention that they had a little sister who died. A character who smiles sadly, trying not to cry, when they bring up a late little sister and who seems irrationally concerned with car safety is already telling the readers the salient points of what happened without need for the character, or prose, to stop and talk about what a good relationship the characters had and the tragic vehicular accident which took the younger’s life (most likely recently). The character who bitterly brings up their dead sibling often and seems dissatisfied when they mention them is already telling the readers that they have unfinished business and bad blood with the sibling, but that they miss them and probably weren’t involved in the death. The character who surprises people by sympathetically giving their first mention ever of a dead little sister when pulling out a kept child’s toy, and who has a sort of grim satisfaction when they mention that it’s what got them into disease research, the character is already telling the readers that their little sibling was terminally ill and was probably given a mercy-kill. And so on. If you have an ex-slave character you don’t need halt a chapter to give a long explanation of the horrors they faced as a slave if you can convincingly have them rub their scarred wrists and mutter a lot number to themselves as they pass some auctioneers in a slaving city during the chapter. After all, most people in reality do not tell their life stories in detail to others – strangers would think they were creepy if they did and non-strangers already know or pick up the important bits as they go.
5. When you are showing how people feel. Show; don’t tell is a rule which is often correctly applied to the portrayal of emotion, but too much showing gives you melodrama. Unless you are deliberately writing a melodrama you do not want to be writing melodrama. This is because of escalation – every time you give an overly lavish description to how someone is feeling, then when you have to describe a stronger feeling you have to describe it even more lavishly. For example, if you write “HOW. DARE. YOU!!!!!!!” he exclaimed, roaring then the response or when you show that character more upset than that you have to add even more to make it clear. While if you write “How dare you?” he roared you don’t have to worry about starting to sound silly when the emotions run higher. Admittedly that’s a slightly exaggerated example, but it works.
VI. Wen u’r beeing kriateive wif speling nd naymes. Funnily enough, naturally occurring languages have shaped themselves to be easily understood – and are what the readers will understand easiest. This also means that you can only go so far with altering or creating names before your readers will no longer be able to pronounce them as you desire because the spelling and pronunciation rules are so alien from the readers’ language. Books like JKR’s Potter series work well with old but real names to give a feeling of a strange new world without tripping up readers terribly, while books like Martin’s A Song of Ice and Fire series manages to get creative with naming without tripping anyone up by changing one letter in otherwise normal names – and by applying a standard rule of spelling of pronunciation, rather than changing things willy-nilly.
And on a final note, if you want further reason to view less is more as a good idea in general: consider how much more readable the later points in this post were than the earlier ones.