Short version: Art and wall decor is a taste-driven category: the buyer cannot spec their way to a purchase the way they can with a pump or a mattress, they browse until something stops them. That changes the playbook. The feed has to speak in subjects, styles and rooms because that is how people search; the structure has to give proven best sellers their own protected home; and scaling has to run on discipline rather than enthusiasm, because taste-driven conversion data is noisy and whipsaws accounts that chase it. We manage art e-commerce and run scaling on a written rulebook: capped budget moves, cooldowns per campaign, and new-customer ROAS as the metric that decides. This guide walks through the whole system.
Why Google Ads works for art and wall decor
Wall decor shoppers search in three registers, and all three convert. Subject and style intent ("abstract canvas art", "japanese wave print", "boho wall decor") is the core: the shopper knows the feeling they want and is looking for the piece. Room intent ("living room wall art", "art above bed") arrives decorating-project-shaped. And gift intent spikes reliably around seasons. Shopping and Performance Max are ideal for the first two because the product IMAGE does the selling in a way no headline can, and Search captures the room and gift framing where a well-built landing page converts a project, not just a product.
The feed speaks in subjects, styles and rooms
The feed decides which taste queries you appear for. Titles built from SKU codes and artist references match nothing; titles built the way people search match everything. For art specifically: front-load subject, style, format and size ("Abstract Ocean Canvas Print, Blue, Large 90x60cm, Framed") because subject plus format plus size is the query; fill color, material, orientation and size attributes since wall decor is filtered visually; keep one clean variant per size and frame option with correct prices; and use custom labels to separate best sellers, new releases and long-tail catalog so campaigns can treat them differently. The general lever is in the product feed most accounts ignore. Images matter double here: the artwork photographed in a room converts the browsing register better than a flat crop.
Campaign structure for an art account
| Campaign | Purpose | Why it is separate |
|---|---|---|
| Brand Search | Capture people searching your name | Cheap, high-converting, must not be absorbed and inflated by PMax |
| Catalog (PMax by style or subject) | Style, subject and room intent demand | Taste-driven browsing behaves differently per style and price band |
| Best sellers (Shopping) | Proven pieces with conversion history | Concentrated data earns tighter targets and protected budget |
| Non-brand Search | Subject, room and gift intent queries | Control and search-term visibility that PMax hides from you |
The best-seller split is the structural move specific to this category. A large art catalog has a long tail of pieces that sell occasionally and a short head that sells constantly; blended together, the head funds the tail invisibly and the account cannot scale what works. Best sellers with real conversion history earn their own campaign, tighter target and protected budget; the long tail lives in catalog PMax where discovery is the job. Brand exclusions on PMax as always, per the truth about Performance Max.
Scaling runs on rules, not enthusiasm
Taste-driven demand is noisy: a piece goes quiet for a week, another spikes on a social mention, and daily numbers invite constant fiddling. The scaling system we run on an art account, written down and followed weekly: budget moves capped at around 10% per change; every campaign gets a cooldown of about two weeks after a change before it can be touched again; and a tROAS change starts the SAME cooldown as a budget change, because a target change is a bid change even though the budget history does not show it. Scaling verdicts come from 7-day AND 14-day windows together - above the floor on one and below on the other is a WAIT, not a scale, because single-window decisions whipsaw budgets. And where campaigns share a budget, a conflict (one qualifies to scale while its budget-mate is mid-cooldown) means hold the shared budget flat and skip both, never split the decision.
The metric that decides is new-customer ROAS where attribution tooling exists, not blended platform ROAS. Art brands accumulate collectors and repeat buyers, and blended ROAS hides that scaled spend often buys customers you already had. Scaling on the new-customer number keeps growth honest.