Distribution and channels

Hotel digital marketing terms: CPC, CPA, ROAS and conversion in plain words

Understand CPC, CPA, ROAS and conversion rate in plain words. Learn what each metric measures, what it hides, and how to judge if a channel is worth keeping.

  • 10 min read
  • AI Otel Editorial

AI draft · awaiting editorial review

Contents

Key takeaways

  1. CPC prices a click, CPM prices a thousand showings and CPA prices a completed action, so the three are different bills, not three ways of saying the same thing.
  2. ROAS is booking value divided by ad spend: a ratio of revenue to cost, not profit, and it says nothing about cancellations or commission.
  3. A conversion rate is only comparable when the top and the bottom of the fraction are defined the same way on both sides.
  4. Retargeting usually reports a better ROAS than prospecting because it advertises to people who were already partway to booking.
  5. The figure that decides whether a channel is worth keeping is what the hotel keeps after cancellation and commission, not what the report calls revenue.

Digital marketing terms are the shorthand an agency report uses for how ad money turned into bookings: CPC is cost per click, CPA is cost per acquisition, ROAS is booking value returned per unit of ad spend, and conversion rate is the share of visitors who book. Each measures one step, not profit.

The monthly report arrives with a row of abbreviations across the top: CPC, CTR, CPA, ROAS. Everyone in the room nods. Nobody asks which bookings were counted, or whether they were still there a month later. This article explains the hotel digital marketing terms an owner or general manager actually meets on a report, in plain words, with what each one cannot tell you. After reading it, you will be able to ask three questions that turn any report into a decision.

What do CPC, CPM and CPA mean?

They are three ways of paying for attention: per click, per thousand showings, and per completed action. The letters are English; the arithmetic works in any currency.

CPC, cost per click, means the hotel pays each time someone clicks the ad. The price is set by an auction, so the same click costs different amounts on different days and in different seasons. CPM, cost per mille, means the hotel pays for a thousand showings of the ad, whether or not anyone clicked. It is used where the goal is being seen, for example a destination campaign before a season. CPA, cost per acquisition, also called cost per action, is the cost of one completed action, and the whole term depends on what the hotel and the agency agreed counts as an action.

The three are linked. Cost per click multiplied by click-through rate, times a thousand, gives the cost of a thousand showings, and a cheap click that never books produces an expensive acquisition. The trap sits in the word "action": a newsletter signup, a booking started and a stay that actually happened are three very different actions, and the CPA falls as the action gets easier.

TermWhat the hotel pays forInvented hotel example
CPCEach click on the ad100 clicks at 1 each costs 100
CPMEach thousand showings10,000 showings at 10 per thousand costs 100
CPAEach completed action2 bookings from that 100 spend is 50 per booking

All figures in the table are invented for illustration. For example, with these invented figures: 10,000 showings at a CPM of 10 cost 100; if 1% click, that is 100 clicks at a CPC of 1; if two of those clicks become bookings, the CPA is 50. If you want a practical starting point for running these campaigns yourself, see how to advertise a hotel: a starter plan for Google and Instagram ads.

What is ROAS and what does it not tell you?

ROAS, return on ad spend, is the booking value an ad produced divided by what the ad cost. A ROAS of 5 means every unit spent came back as five units of booking value, before anything else is subtracted.

It is a ratio, not an amount and not profit: it compares two numbers and stops there. Which booking value was used matters: the room rate at the moment of booking, the rate after discount, or the money actually collected after the stay. Three choices, three different ROAS figures on the same campaign.

Cancellations stay in the month they were booked. A booking cancelled two weeks later still lifts the ROAS of the month it was made, and the report rarely removes it. Commission and the cost of serving the stay are outside the ratio: a ROAS comfortably above 1 can still describe a booking that lost money once the channel's commission and the room's running costs are counted. Size matters too. A small spend with one lucky booking shows a spectacular ROAS; a large spend with many bookings shows a modest one, and the second campaign may be the better business.

For example, with invented figures: 1,000 spent, 6,000 of bookings at the rate booked, of which 1,500 later cancels and the rest arrives through a channel that takes a commission. The reported ROAS is 6; what the hotel keeps is a smaller number that the report does not show.

Watch outA high ROAS is not a profit signal

Before accepting it, ask three questions: which bookings were counted, were cancelled bookings removed, and is the value the rate booked or the money collected. For the fuller comparison, see distribution cost per booking: comparing channels by what the hotel keeps.

What is conversion rate on a hotel website?

Conversion rate is the share of visitors who complete the action the hotel wants, usually a confirmed booking. It is a fraction, so it means nothing until both the top and the bottom of the fraction are defined.

The top of the fraction can be a booking confirmed, a booking started, a room held or a request sent; the same traffic produces four different rates. The bottom can be sessions or users, all visitors or one source market, all devices or desktop only. Mixing them between two reports creates a change that never happened.

Look-to-book, the ratio of availability checks to confirmed bookings, survives seasonality better than a raw visitor rate, because it compares people who were already shopping. A hotel booking is also a considered purchase: visitors compare, leave, and return on another device days later, so a single-session rate undercounts the real path to booking. Micro-conversions, such as an availability check or an email signup, are useful for judging a page. They are misleading when used to judge a channel.

ChecklistBefore comparing two conversion rates, confirm all of the following

  • Same numerator: confirmed bookings, not started bookings
  • Same denominator: sessions, not a mix of users and sessions
  • Same date range and time zone
  • Cancelled bookings removed or kept, stated either way
  • Bot and internal traffic filtered the same way
  • Same device and country filters

A change of definition can look exactly like a change in performance, so the definition belongs in the report header, not in a footnote.

What do impression, reach, click and CTR mean?

An impression is one showing of an ad, reach is the number of different people who saw it, a click is a tap through to the hotel's page, and CTR is clicks divided by impressions.

Impressions count showings, not people: one person shown the same ad five times produces five impressions and a reach of one. Frequency is impressions divided by reach. A high frequency with a falling click rate usually means the audience has seen enough of the ad, not that the ad is weak.

CTR, click-through rate, is clicks divided by impressions, written as a percentage. It measures the ad's ability to earn a tap, nothing further. It is not comparable across formats: a search ad answers a query the person just typed, while a display banner interrupts a page they were reading. Compare like with like or not at all.

A click is not a visit. Some clicks are accidental, some leave before the page loads, and some are blocked by the browser, so the analytics number will be lower than the ad platform's number.

NoteAd platform clicks and website sessions rarely match

The difference comes from accidental clicks, blocked scripts and consent choices. Reconcile the trend across months, not the exact number on one day.

What is retargeting, and what is a lookalike or a segment?

A segment is a group of people defined by a shared behavior, retargeting advertises to a segment that already visited the hotel, and a lookalike is a new audience built to resemble an existing one.

A segment can be visitors who checked availability for a date and did not book, past guests from one source market, or people who opened the last three emails. Retargeting shows ads to that segment later, on other sites or in social feeds. The window matters: intent fades, so a 7-day window and a 90-day window are different campaigns with different results.

A lookalike, sometimes called a similar audience, is built by the platform from the pattern of a seed list. Its quality depends entirely on the seed, so a list of past guests and a list of newsletter signups produce very different audiences. Retargeting usually reports a better ROAS than prospecting because it advertises to people who already know the hotel and were partway to booking; it takes credit for demand that partly existed already.

Retargeting depends on tracking people across sites, which involves personal data. The hotel carries the duty for notice and consent, and the exact wording and rules should be confirmed with a legal adviser. A frequency cap is part of the setup, not an optimization later: the same guest shown the same ad twenty times is a cost with no return.

What is the difference between organic, paid, brand and non-brand search?

Organic is the free listing, paid is the bought one, brand search contains the hotel's name, and non-brand search does not. The four boxes answer different questions and belong in separate rows of the report.

Organic results are the ones the search engine chose. The hotel does not pay per click; it pays in content, page speed and time. Paid results are the ads above or below the listings. Cost is per click and the position is auctioned, so the same keyword can cost more in high season.

Brand search means the person already typed the hotel's name. A high conversion rate here is expected, because the decision was largely made before the click. Non-brand search means the person typed a destination, a date or a need. This is where new demand is won, and where the cost per booking is normally higher. Blending brand and non-brand into one average hides both, so ask for the split before judging a campaign.

Metasearch and hotel price listings sit between the two: the hotel's rates appear next to other hotels' rates for the same dates, and the traveler compares in one place. How those listings are priced and booked is covered in Google hotel prices: Hotel Center, free booking links and hotel ads.

What do these numbers leave out?

They leave out cancellations, commission, the cost of serving the stay, and the guest who books again next year. What remains after those is the only figure that decides whether a channel is worth keeping.

Booked versus stayed is the first gap. A report dated by booking date and a report dated by stay date describe different months, and comparing one with the other produces false conclusions. Ask for the cancellation rate of the bookings attributed to each channel, not the rate for the hotel overall; the two can move in opposite directions.

A channel with a high ROAS and a commission can leave less than a channel with a lower ROAS and no commission, so the comparison has to be made after the commission, not before. A first booking that later produces two direct bookings is worth more than its own value, and no ad report shows this without a guest identifier the hotel controls. Cost per stay, not cost per booking, is the sharper number: divide the spend by the stays that actually happened, and the ranking of channels can change.

ChecklistQuestions to ask when the report arrives

  • Which bookings count: made, confirmed or stayed?
  • Are cancellations removed, and in which month?
  • Is the value the rate booked or the money collected?
  • Is commission included?
  • Are brand and non-brand split?
  • What is the cost per stay, not per booking?
  • Is the period based on booking date or stay date?

Keep one table every month with the same columns: spend, bookings made, bookings cancelled, stays, money collected, commission, net. A stable table beats a clever one. The wider vocabulary around channels and contracts is in hotel distribution terms explained: contracts, rates, channels.

What to do next

Ask for one extra column before the next report: what the hotel kept. Take the channel that spends the most, request its bookings split into made, cancelled and stayed, subtract commission, and set that number beside the same number for direct bookings. The vocabulary above earns its place only if it lets you ask that question in the meeting.

Frequently asked questions

Is a ROAS of 4 good?

There is no universal threshold, and any figure quoted as one should be treated with suspicion. A ROAS has to be read against the hotel's own margin, the channel's commission and the share of those bookings that survive to the stay. Ask what value was counted and whether cancellations were removed before comparing two campaigns.

What is the difference between CPA and cost per booking?

CPA is the general term for the cost of one completed action, and a booking is one possible action. Cost per booking is therefore a CPA with the action defined as a booking. If the action is defined as a stay that actually happened, the correct name is cost per stay, and the number is higher.

Why does the ad platform show more clicks than my analytics show visits?

Accidental clicks, blocked scripts, consent choices and redirects all remove sessions between the ad and the analytics tool. The two systems also count at different moments. Reconcile the trend across months rather than chasing an exact match on a single day.

Should a hotel run retargeting at all?

It can be worth it when the hotel already has website traffic and a lawful basis for the tracking it requires. Keep the window short, cap the frequency, and judge it against the bookings it produced rather than against the better ROAS it will almost always show. Confirm the data duties with a legal adviser.

What is a good conversion rate for a hotel website?

No figure can be given without knowing the definition, the traffic mix and the source markets, so treat any published benchmark as someone else's number. Compare the hotel with itself: same numerator, same denominator, same filters, month against month. A definition that changes between reports will produce a change that never happened.

AI Otel EditorialEditorial team

Drafted with AI assistance from AI Otel’s published fact base. Not yet reviewed by an editor.

For hotels: OtelCiro.

OtelCiro, AI Otel’s hotel growth and consultancy arm, is operating today.

Talk to OtelCiroRequest a briefing