Lookbook
A lookbook is a curated set of images that presents a clothing collection as complete styled looks rather than as individual products, used to communicate the direction of a season to buyers, press and customers.
The unit of a lookbook is the look, not the garment. A product page sells a jacket; a lookbook shows the jacket with the trousers, the shoes and the styling that tell you what the brand thinks the jacket is for. That is why the same piece can appear in three looks and read as three different propositions.
What makes a lookbook work is consistency across the set. One model, one lighting setup, one grade and one background language, held across every frame, is what makes twenty images read as one collection instead of twenty photographs. Practitioners call the shared treatment a style anchor, and holding it is more important to the finished set than any single hero image.
Consistency is also the part that is hard to reproduce with generative tools. A written style anchor pasted into every prompt will hold the mood and the grade, but it will not hold the same face or the same garment from frame to frame, and a set where the model changes subtly between looks reads as a mistake rather than as a collection.
The production shape has changed more than the format. A lookbook is a matrix: fix the model, the scene and the finishing, then vary the garment. Setting those once and iterating the clothing against them is a closer match to how the shoot actually works than describing the whole picture again for every look, and it is what makes a full collection viable outside a studio day.
Frequently asked questions
What is the difference between a lookbook and a catalog?
A catalog is organised around products and exists to be bought from, so each image isolates an item. A lookbook is organised around styled looks and exists to communicate direction, so images show complete outfits in a consistent world.
How many looks should a lookbook have?
There is no rule, but a seasonal drop is commonly presented as ten to twenty looks. The constraint is usually coherence rather than count: a shorter set that holds together beats a longer one that does not.
What keeps a generated lookbook consistent?
Fixing the variables that should not change. A roster model carrying a character sheet holds the face far better than describing the same person repeatedly, and holding one scene and one finishing look across the set does the same for the treatment.