Custom Candle Trio, rebuilt | P.F. Candle Co.

Build Your Own Candle Trio

15% off any three candles, any size. Three classics come to $72 and drop to $61.20 in cart. You save $10.80.

Why this converts

The trio page, today vs proposed

Everything on the left is from pfcandleco.com/products/custom-candle-trio as of 10 Aug 2026. Everything on the right is live on this page; try it above.

Choosing a scent

Each card gives three words and a price. Nothing on the card opens; the scent notes live on 18 separate product pages.

The full scent story sits one tap away on every card: the mood, then top, heart and base notes, without leaving the builder.

First impression

The tray starts at $0.00 with a greyed-out button. The 15% lives in one line of fine print.

The page opens anchored at $61.20 with the $10.80 saving on screen before the first click, counting up as you pick.

Finding your way

Three size tabs split the range, so two thirds of it is invisible at any moment.

One grid, filtered by scent family, size picked on the card. The main nav already sells by scent; now the builder does too.

The range on show

Amber & Moss, the original best seller, is in stock on its own page but missing from the builder. Sold-out scents vanish; the HI-FI tab is down to one candle today.

The whole scent library stays visible with honest availability. Sold-out jars say back soon instead of disappearing.

Momentum

Progress is a thin bar and a line of small text above three grey boxes.

A trio tray fills jar by jar, the blend profile updates as you pick, and one tap completes the trio with best sellers.

Learning from shoppers

One fixed layout, no measurement of what candle buyers respond to.

Two layouts under a live 50/50 test, decided by attach rate, completion and AOV, with scent-mix data as a bonus for planning restocks.

The experiment

How we run it as a real A/B test

The toggle at the top of this page is a demo control. In production, Kitenzo assigns each visitor one variant and holds it for their whole session.

Split

Visitors landing on the trio page are assigned A or B once, 50/50. Nobody ever sees both, so the numbers stay clean.

Same engine

Both variants run the same Kitenzo rules, cart grouping and Cart Transform pricing. Checkout and the 15% discount are identical; only the shopping experience differs.

Measure

Every pick, swap, completion and add-to-cart is an event. The dashboard reads each metric per variant so the winner is a number, not an opinion.

Ship the winner

Promote the winning layout to 100% of traffic in one click. Keep the other as a seasonal re-test, for gifting season, for example.

Builder attach rate

Of visitors who land on the page, how many start a trio.

Trio completion

Of shoppers who pick one candle, how many reach three.

Bundle AOV

Average trio value; do shoppers trade up to large and HI-FI jars?

Scent mix

Which scents actually get picked together; feeds restocks and future sets.

Variant A asks shoppers to build from an empty tray. Variant B starts them from a full one. Retail data argues both ways; this is exactly the kind of question a test settles better than a redesign debate.