A lot of writers treat comp titles as a checklist exercise: name three books that sold well in your genre, paste them into the query, done. That misunderstands what a comp set is for. Comps are a positioning statement — a sentence about where your book sits on the shelf, said in the shorthand of other book titles. Chosen well, they do real work. Chosen lazily, they actively hurt you. Here is a short field guide.

What comps are actually doing

A comp set is talking to three different audiences at once, and it has to satisfy all of them.

It tells an agent or editor that you understand your own market — that you know where your book lives and who it sits beside. A query with sharp comps signals a writer who has read in their lane and thought about positioning, which is a quiet but powerful credibility signal.

It tells a reader what kind of experience to expect. "For readers of X and Y" is a promise about tone, pace, and feel, and readers use it to decide whether your book is for them.

And it tells the retail algorithms where to file you. On the major stores, the books you sit beside shape who gets shown your cover. Comps aren't only a pitch document; they're an input to discovery.

The rules of a good comp set

Recent. Aim for the last three or four years. A comp from a decade ago tells an agent the market has moved on without you. Recency signals that you are reading what is selling now, not what sold when you started writing.

Same audience, same shelf. The point of a comp is overlap of readership, not plot. A book about a heist and a book about a wedding can be perfect comps if they hit the same reader with the same tone. Comp to the audience, not the logline.

Honest about scale. This is the one most writers get wrong, so it gets its own section below.

Specific. "Bestselling fantasy" is not a comp. A named title is. Vagueness reads as not having done the work.

The mega-bestseller trap

The instinct is to reach for the biggest name in your genre. Resist it. Comping your debut to a generational phenomenon does two things, both bad. It invites a comparison you cannot win — no debut survives being measured against a cultural juggernaut. And it tells an agent you don't understand the market, because those books are outliers, not categories. Nobody is acquiring "the next" anything; they are acquiring a book that fits a real, sized audience.

The better move is to comp slightly below the megahit, to the strong-performing, well-reviewed books one tier down. They prove the audience exists and is buying, without making a promise your book can't keep. A solid mid-list comp that genuinely matches is worth more than a blockbuster that doesn't.

The "X meets Y" formula, used well

"X meets Y" is a useful tool and an easy crutch. It works when the two halves are genuinely surprising together and the collision says something true about your book — one supplies the world or genre, the other the tone or twist. It fails when it's two random bestsellers stapled together to sound impressive. If the formula doesn't actually describe your book to someone who knows both titles, drop it.

How many, and how to find them

Three to five. Fewer than three is too thin to triangulate a position; more than five dilutes the signal and starts to look like you are guessing. For the mechanics of finding good candidates — recent genre reads, retailer also-boughts, Goodreads shelves, category bestseller lists — we wrote a separate guide: how do I find comps for my book.

Where AI helps, and where it doesn't

The research half of comp work — surfacing recent titles in your space, pulling who reviewed them, spotting also-boughts — is slow, and it is exactly the kind of pattern-matching that AI comp curation speeds up dramatically. It can hand you a long list of plausible candidates in the time it takes to make coffee.

But the choice is still yours, and it is a judgment call, not a search result. Which five position the book honestly? Which flatter it without lying? Which collision of comps tells the truest story about what you wrote? A model can find candidates. Only you can decide what your book is standing next to. That decision is the positioning statement, and it is worth getting right.