Hotel Marketing Metrics That Actually Matter.

Direct booking share

Formula — direct-channel room-nights ÷ total room-nights, trailing 12 months.

This is the single number that tells you who owns your guest relationship: you or the OTA. Across independent hotels, online travel agencies took roughly 63% of bookings in 2025 (Cloudbeds), leaving independents with a direct share in the high thirties. Branded hotels, with their loyalty machines, keep OTA reliance closer to 35% (Phocuswright). The healthy target that keeps surfacing across operators isn’t 100% direct — it’s 40–50%, with OTAs kept on as a paid billboard rather than a landlord.

Why it matters: every point of direct share you win back is a point you stop paying commission on, and a guest whose email address you now own. It’s the metric that turns a marketing conversation into a margin conversation, which is why it belongs at the top of the dashboard, not buried in a channel report. For the full picture of where boutique properties actually sit, see The Direct Booking Benchmark.

The goal was never zero OTA. It’s owning enough of the channel that owns the guest.

Website conversion rate (and its more useful cousin)

Formula — completed direct bookings ÷ unique website sessions × 100.

Conversion rate is the multiplier on every dollar you spend driving traffic. Double your conversion and you’ve effectively doubled your ad budget for free. Average hotel website conversion runs 2.2–3.9%, with the best boutique sites clearing 5% (STR and multiple 2026 sources). If you’re at 1.8%, more traffic is water poured into a leaking bucket.

But the site-wide number hides the real story, so track a second version: booking-intent conversion — bookings divided by the sessions that actually reached your booking engine. That runs far higher, typically 8–15%, and it separates two very different problems. Low site-wide conversion but healthy booking-intent means your site isn’t guiding people to the engine. Low booking intent means the engine itself is losing them, and roughly 62% of people who open a hotel booking engine never finish. One is a content and navigation problem; the other is a checkout problem. You can’t fix what you haven’t separated.

One more cut worth watching: mobile versus desktop. Mobile is now around 62% of hotel website traffic but converts at roughly half the desktop rate. If your gap looks like that, the villain is usually mobile UX, not your pricing.

Cost per acquisition (a.k.a. cost per direct booking)

Formula — total direct-channel acquisition cost ÷ direct bookings — where cost includes paid media, booking-engine fees, tooling and management.

This is the metric almost nobody publishes and everyone should. “Direct is cheaper than OTA” is only true if your fully loaded cost per direct booking comes in under the effective OTA commission you were trying to escape. That commission runs 15–25%, and up to ~30% once you’ve stacked on promotional placements. Well-run metasearch and Google Ads campaigns deliver direct bookings at a CPA of 8–14% of booking value (comfortably under OTA cost), but a neglected campaign pointed at a site that doesn’t convert can quietly cost you more than the commission. At that point, you’ve simply swapped a commission bill for an ad bill and called it a strategy.

The discipline: express CPA as a percentage of booking value and hold it against your real OTA commission. If direct CPA is 12% and OTA commission is 20%, direct is winning. If they’re both 20%, your infrastructure, not your ad spend, is the problem. The Benchmark data puts portfolio-median CPDB at 9%.

Direct isn’t a discount. It’s an investment that only pays back when the booking experience is built to convert.

Return on ad spend (ROAS)

Formula — revenue attributed to ads ÷ ad spend.

A $1,000 spend that drives $8,000 in direct bookings is an 8:1 ROAS. ROAS is your primary efficiency read on paid, and for independent hotels a target of 8:1 or better is realistic on metasearch and Google Hotel Ads. But a headline ROAS number lies by omission unless you split it two ways.

First, brand versus non-brand. Campaigns bidding on your own hotel’s name convert cheaply and post gaudy returns (8–13:1 and above is common), but a chunk of that is demand you’d likely have captured for free through organic. Non-brand campaigns, which create new demand from people who don’t yet know you, run a more honest 2–5:1. A blended ROAS that looks spectacular is often just brand campaigns flattering the average.

Second, impression share. You can inflate ROAS by only ever bidding on the cheapest, highest-intent traffic — and in doing so, sit out most of the auction while OTAs capture the demand you skipped. A high ROAS with low impression share isn’t efficiency; it’s under-investment. Read them together.

Blended CAC vs. channel CAC

Formula — blended = total acquisition spend ÷ all new bookings; channel = same math, run per channel.

The distinction trips up more operators than any other metric here. Blended numbers feel reassuring and hide everything that matters: a cheap, brand-heavy channel can subsidize a wasteful one and leave the average looking fine while you quietly overpay for half your bookings. Manage to blended CAC and you’ll never know which lever to pull. Manage to channel CAC and you can move budget from the 5:1 channel to the 9:1 channel with confidence. Blended is the number you report; channel is the number you act on.

Average booking value by channel

Formula — net revenue ÷ bookings, calculated separately for each channel.

Not all bookings are equal, and the channel they come through predicts their worth. Direct bookings averaged around $516 in revenue in 2025 versus roughly $312 on the OTAs (SiteMinder) — direct guests book higher room categories, stay longer, and add extras, because they arrived through your storytelling rather than a price-sorted list. This is the quiet argument that reframes the whole direct-versus-OTA math: you’re not just saving commission, you’re winning a more valuable guest.

Guest lifetime value (and repeat-direct rate)

Formula — average booking value × average bookings per guest × average retained years.

A single booking dramatically undervalues a guest you can bring back. Direct bookings run about 12.5% more profitable than OTA even after ad costs (Kalibri Labs), and loyalty/repeat guests are nearly twice as profitable again — they cost almost nothing to re-acquire and they spend more. The practical proxy if full LTV is too heavy: track your repeat-direct rate, the share of direct bookings from guests you’ve hosted before. It’s the clearest sign your marketing is building an asset, not just renting traffic.

RevPAR — the number the owner actually cares about

Formula — ADR × occupancy (equivalently, room revenue ÷ available rooms).

A caveat: RevPAR is a revenue-management metric, not strictly a marketing one — it’s shaped by pricing and inventory as much as by demand generation. But it’s the number your owner or GM lives by, so it’s the “so what” your marketing metrics should ladder up to. When you improve direct share, conversion, and channel efficiency, RevPAR is where it’s supposed to show up. Reporting your marketing wins in the language of RevPAR is how marketing earns its budget.

Email: revenue per send, not open rate

Formula — revenue attributed to email ÷ emails sent (plus email-attributed booking conversion).

Email is one of the highest-return channels a hotel owns — average hotel email marketing returns around $38 for every $1 spent, and well-built lifecycle automations can reach $100–200. Hotel email converts to bookings at roughly 2%, and automated, behavior-triggered emails generate on the order of 22× more revenue per send than one-off blasts.

Which is exactly why you should stop leading with open rate. Apple’s Mail Privacy Protection now counts an email as “opened” when its server preloads the tracking pixel — no human required — inflating hotel open rates by an estimated 10–30% and rendering the metric close to meaningless. Judge email on revenue per send, click-through, and booking conversion. Those tie to money and can’t be faked by a privacy feature.

None of these are worthless — some are useful diagnostics upstream. They become dangerous when they’re mistaken for the goal.

Impressions and reach. They measure how many screens you touched, not how many rooms you sold. Reach with no downstream conversion is a lighthouse in an empty ocean.

Followers and likes. A vanity leaderboard. A 200,000-follower account that drives no bookings is worth less than a 4,000-person email list that does.

Raw traffic / sessions. More visitors is only good if they convert. Chasing sessions without watching conversion is how properties end up paying to fill a leaking bucket.

Email open rate. Broken by Apple MPP, as above. Track it only as a loose trend for non-Apple segments, never as a target.

Click-through rate in isolation. A high CTR that doesn’t produce bookings just means you wrote a good hook and a bad landing page. CTR only matters chained to conversion.

Bounce rate on its own. A high bounce on a page that answered the visitor’s question and sent them to the booking engine isn’t a failure. Context or nothing.

If a number can go up while your revenue stays flat, it’s a diagnostic at best and a distraction at worst.

The metrics that matter aren’t a list, they’re a chain. Marketing gets more of the right people to the site (direct share, channel CAC, ROAS); the site turns them into bookings (conversion rate, booking-intent conversion), those bookings are worth something (average booking value by channel); and the best guests come back (repeat-direct rate, LTV), all of it landing, eventually, in RevPAR.

Read that way, most “underperforming marketing” reveals itself as a single weak link. Traffic is up but bookings aren’t? Conversion. Bookings are up but profit isn’t? Channel CAC. Everything’s up but nobody comes back? Retention. You don’t need forty metrics. You need the six or seven that form the chain, watched together, so you can see which link is breaking.

Measure the chain, fix the weakest link, ignore the noise.

Ranges compiled from Cloudbeds, Phocuswright, SiteMinder, STR, Kalibri Labs, and current 2025–2026 hospitality marketing data. Reference points, not targets.

They’re optimizing a dashboard full of numbers that were never going to move revenue, while the six that would sit unwatched.

If you’re not sure which of your numbers are signal and which are noise, or you’re watching your conversion rate climb without knowing what it’s costing you to get there, that’s the work: building the measurement layer that shows you the chain, then fixing the link that’s actually breaking.