A single attach rate, quoted without its definition, is one of the easiest ways to be wrong about ancillary performance while sounding precise. Two teams can run the same merchandising, sell the same extras to the same travellers, and report numbers that differ by a factor of three. Neither of them is lying.
In short
Attach rate is the share of bookings that gain at least one paid extra. The number only means something alongside its denominator, because bookings, passengers and segments each produce a different figure from identical sales. Define the denominator first, compare only like with like, and always read revenue per booking beside it.
What the number is trying to tell you
Attach rate asks a narrow question. Of the trips you sold, how many ended up carrying something else that was paid for? It is a merchandising measure rather than a traffic measure, and that is its value. Impressions, clicks and page views all rise when you put an offer in front of more people. Attach rate only rises when someone decides the offer was worth taking.
That makes it the right diagnostic for a post-booking programme and the wrong one to compensate anybody on. It says whether relevance is working. It says nothing at all about how much money arrived.
The denominator decides the number
There is no standard denominator in this industry, which means the first question to ask about any attach rate is what sits underneath the division line.
| Denominator | One unit is | Effect on the number | Suits |
|---|---|---|---|
| Bookings | One confirmed reservation, whatever its size | Highest of the three trip measures | Cart and merchandising performance |
| Passengers | One traveller on that reservation | Falls sharply once family bookings are in the mix | Products every traveller decides on individually |
| Segments | One flight within the itinerary | Lowest, and lower again on connecting trips | Products priced flight by flight |
| Eligible bookings | Only the trips the product could apply to at all | The most flattering number available | Judging the offer rather than the audience |
Take a family of four flying a return trip with one connection in each direction. That is one booking, four passengers, and four flights. Sell them a single bag and you have attached to one booking in one, one passenger in four, and one flight in four. Nothing about the sale changed. The number moved from all of it to a quarter of it.
The eligible-bookings denominator deserves particular suspicion. It is a genuinely useful diagnostic, because it separates a product nobody wants from a product almost nobody could have bought. It is also the number that gets quoted in a board pack without its qualifier, at which point it is simply wrong.
Comparing channels without fooling yourself
Most attach rate comparisons between two channels are really comparisons between two trip mixes. A channel weighted towards long haul, or towards travellers on fares that already include a bag, or towards groups rather than solo travellers, will report a different rate for reasons no merchandising change can touch.
Four things have to match before a comparison means anything: the denominator, the eligibility filter, the measurement window, and the definition of a sale. That last one catches people out, because an item added to a cart and an item paid for are different events, and dashboards do not always agree about which one they are counting.
When those cannot be aligned, compare each channel to its own past instead. A channel improving against itself is a real finding. A channel beating another channel usually is not.
Read it as a funnel, not as a rate
One number cannot tell you what to fix. Three can.
Split the rate into eligibility, exposure and purchase. How many bookings could have carried the product at all? Of those, how many were actually shown it? Of those, how many bought? A weak result in the first band is a content or coverage problem. A weak result in the second is a placement or timing problem. A weak result in the third is a price, clarity or trust problem. These have almost nothing to do with each other, and a single blended rate hides which one you have.
A high rate on a cheap item can be worth less
The reason attach rate should never be the only number on the page is that it is blind to value.
Revenue per booking is the number that survives that comparison, and the two belong together. Attach rate tells you whether the offer was relevant. Revenue per booking tells you whether relevance was worth having. A programme optimised on the first alone drifts towards cheap items that everybody takes, which is a real and quiet way to lose money while every chart points upwards.
The practical habit is to report the pair, always, with the denominator written next to them. It is a small discipline that removes most of the arguments in the room, and it is the difference between a metric that guides a decision and a metric that decorates a slide.
Frequently asked questions
Which denominator should we use for attach rate?
Use the unit the product is sold in. Seats are decided by each traveller, so passengers is the honest denominator. A bag is usually decided once for a whole party, so bookings fits better. Whichever you pick, publish it next to the number.
Is attach rate the same as conversion rate?
No. Conversion rate measures whether a step completed, such as a checkout that finished. Attach rate measures whether the booking ended up carrying an extra. A page can convert well and still attach badly if the offer was never relevant.
Can attach rate fall while revenue rises?
Easily, and it is one of the most common ways a healthy quarter looks like a bad one. Adding a high value item that suits fewer travellers lowers the share of bookings that attach anything while raising revenue per booking.
How should two sales channels be compared?
Only on the same denominator, the same eligibility filter, and the same measurement window. Even then, differences in trip mix explain more than merchandising quality does, so compare each channel against its own history first.