Benchmarking your funnel against the sector

The slide that does the most damage in a travel marketing review carries two numbers: yours, and "the sector". A regional hotel group put its 2.1% booking conversion next to a published 3.0% sector median and asked its agency for a recovery plan. The arithmetic was right and the conclusion was wrong: the two figures counted different denominators, over different months, on a different traffic mix. Before you fund a fix, you have to decide whether a gap is measurement, mix or performance, in that order. That is a leadership call, not an analyst's.

Which sources are actually comparable

Two disclosures worth making out loud when you quote these: STR sells hotel benchmarking as its product, and customer data platformscustomer data platformsSoftware that unifies customer data from every source into one persistent profile that marketing, sales and service teams can act on.View full definition → such as Segment sell the tracking layer whose event definitions decide what your own numbers say. Neither fact disqualifies them. It does mean their published medians describe the operators who buy from them.

The most common break is the denominator. A standard analytics session closes after 30 minutes of inactivity, so a guest who compares three properties across a lunch hour produces two sessions and one booking. Session-based and user-based conversion on identical traffic can differ by a third. Your booking engine counts confirmations; the ad platform counts modelled conversions inside its own attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → window. Three defensible conversion rates, and none of them is the one in the benchmark table.

STR's habit is worth borrowing even if you never buy a STAR report: it publishes indexes against a comp set the hotel itself names, where 100 means you took fair share of that set. Marketing benchmarking almost never has that discipline. Building it yourself, by naming eight to twelve properties or operators you genuinely compete with for the same guest, does more for comparability than any purchased median.

Normalise for mix before you accept the gap

The hotel group's 2.1% was a blend. Non-brand metasearch and paid social carried 70% of sessions at 1.2%; brand and direct carried 30% at 4.2%.

blended     = 0.70 x 1.2% + 0.30 x 4.2% = 2.10%
panel mix (40% non-brand / 60% brand):
normalised  = 0.40 x 1.2% + 0.60 x 4.2% = 3.00%

At the panel's traffic mix, with the group's own per-segment rates untouched, it lands exactly on the median. There is no conversion problem. There is a portfolio buying more upper-funnel traffic than the panel does, which someone chose on purpose and should defend rather than apologise for.

Reweighting works on any dimension the source discloses: lead time, party size, corporate versus leisure, device, market. If the source will not state its mix, you cannot normalise, and the figure is gossip, not a target.

Season, and the trap of the trailing window

Easter 2024 fell on 31 March, Easter 2025 on 20 April. A UK short-break operator comparing March to March moved an entire holiday peak from one side of the line to the other, and its conversion rateconversion rateThe percentage of visitors or prospects who complete a desired action (purchase, sign-up, contact form), calculated as conversions divided by total opportunities.View full definition → "collapsed" by a fifth without anything changing. Trailing 90-day dashboards do this quietly every quarter as they roll out of peak and into shoulder.

Three rules hold the season constant. Compare the same calendar weeks, not the same day counts. Hold the booking window: conversion in the final week before travel runs far above conversion 90 days out, so a year weighted towards late bookings flatters the same website. And check whether the benchmark is struck on stay date or booking date, because a resort selling summer in January will look broken on one basis and healthy on the other.

When you really are below median

Work the three explanations in cost order. Measurement is cheapest: reconcile booking engine confirmations against analytics, expecting a gap in the high single digits to low double digits from consent choices and blocked tags, and settle whether you are quoting sessions or users. That is weeks of work, not a quarter. Mix comes next, and needs no new spend at all. Only then is it performance, at which point you locate the leak using the stage transitions the funnel lesson defines and the pre-booking signals the engagement lesson owns, compare cost using the fully-costed acquisition figure rather than media spend, and set the value side against the multi-year repeat value the lifetime-value lesson models.

Sometimes below median is the right place to sit. A resort at $900 a night with a 60-day booking window may convert at 1.4% while a city-centre property converts at 4%, and the resort's economics tolerate it because one booking is worth several. Half of any sample sits below its median by construction. Writing the median into next year's objectives sets a ceiling for half your portfolio and a target nobody can miss for the other half.

Failure modes leaders own

• Amputation. Cut non-brand prospecting and both conversion rate and cost per booking improve on paper while bookings fall two quarters later. The brand-defence line then claims credit for demand it did not create.
• Definitions drifting to meet the target. If the median is in a bonus plan, someone will re-scope an event or move a stage boundary. Freeze definitions for the year, version them, and date every change.
• Comp set vanity. A 40-room independent measured against branded chain panels will always look short on retention, because the comparison includes a loyalty currency it does not have.
• Single-source dependence. One vendor median, quoted twice in a board pack, becomes fact by repetition.

🎬 [VIDEO: "How OTAs and Direct Booking Compete for the Same Guest" - youtube.com/results?search_query=OTA+vs+direct+booking+hotel+marketing - search for recent explainer videos on hotel distribution channel economics and commission structures]

Building your own benchmark file

Keep one page per metric: the definition, the denominator, the source, the mix it assumes, the season basis, and the date pulled. Anything quoted in an executive review without those six lines gets sent back. Then triangulate:

1. Public investor reports, for direction of travel on marketing efficiency and direct share
2. Industry research: Skift Research publishes travel-specific marketing and distribution data, and Google's Think with Google Travel hub tracks consumer booking behaviour trends
3. Your own history, on stable definitions, which usually beats any external figure for deciding whether this quarter is genuinely worse

Key Takeaways

• Diagnose a gap in cost order: measurement, then mix, then performance. Most gaps close on the first two.
• A benchmark you cannot normalise is not a benchmark. If the source hides its denominator and its traffic mix, do not put it on a slide.
• Reweighting your own segment rates to the panel's mix often erases the whole gap, as 2.1% became 3.00% above.
• Season needs calendar weeks, a held booking window and an explicit stay-date or booking-date basis; Easter alone can move a quarter by a fifth.
• Borrow STR's comp set logic for marketing: name who you actually compete with for the guest, and index against them.
• The sector median is a diagnostic, not an objective. Written into a bonus plan it becomes a ceiling and an invitation to redefine the metric.