Most of the paid media advice aimed at independent hotels is about catching people who are already shopping. Search ads on your own name, metasearch bids on the date a guest picks your property, retargeting the abandoned booking. All of that is demand harvesting. It works, and you should do it. But it has a ceiling: you can only capture intent that already exists.
Demand Gen is the other half of the equation. It is Google’s answer to the kind of visual, scroll-stopping advertising you used to only get on Meta, and it runs across YouTube, Discover, and Gmail. For a boutique property, it is the cheapest way I know to put your place into someone’s head before they have decided where to go. Today I want to walk through exactly how I set these campaigns up for hotel clients, what creative actually works, and the part everybody gets wrong: measuring it honestly so you do not either fool yourself or kill a good campaign too early.
What Demand Gen actually is (and is not)
Demand Gen replaced the old Discovery campaign type and absorbed a lot of what used to live in standalone YouTube campaigns. The inventory is the good stuff: the YouTube home feed, in-stream and Shorts, the Discover feed on people’s phones, and the Gmail promotions tab. It is image-and-video first. There is no keyword targeting. You feed it creative and audiences, and Google’s models decide who sees what.
That last bit matters. Demand Gen is not a search campaign with pretty pictures bolted on. You are not bidding on intent. You are renting attention from people who fit a profile and trusting the system to find the ones most likely to care. That is a different mental model, and if you judge it by last-click ROAS the way you judge a brand-search campaign, you will conclude it does not work and you will be wrong.
The honest framing: search ads capture the guest who already wants to come. Demand Gen plants the idea in the guest who does not know your town exists yet. Different jobs, different yardsticks.
If you have not yet sorted out the campaigns that capture intent, start there first. A Demand Gen campaign pouring interest into a funnel with no book-direct conversion path or a broken metasearch setup is just paying to send warm traffic to the OTAs. Build the catch net before you turn on the firehose.
Why this fits independent and boutique hotels specifically
Big-box chains do not need Demand Gen the same way. Their brand does the destination work. When someone thinks “I need a hotel in this city,” a Marriott or Hilton is already in the consideration set. You, the independent, are not, unless you put yourself there.
Boutique properties also have an unfair advantage in this format: you are visual. A converted 1920s firehouse, a courtyard pool framed by bougainvillea, a rooftop at golden hour, a tasting menu plated like art. That stuff stops a thumb. Generic chain rooms do not. Demand Gen rewards distinctive imagery, and distinctive is exactly what an independent hotel has and a chain does not.
The whole point of Demand Gen for an independent is to manufacture the brand recognition that chains buy with scale. You are using lifestyle creative to make a stranger think of your property by name later, so that when they finally search, they search for you instead of a generic phrase the OTAs win.
That last sentence is the strategic core. If you want to understand why winning your own name in search matters so much, I wrote a whole piece on why your hotel ranks below the OTAs for your own name. Demand Gen is upstream of that fight: it creates the branded searches in the first place.
How I structure a hotel Demand Gen campaign
Here is the build I actually use. It is not complicated, but the order matters.
1. Audiences: lookalikes first, interests second
The single best signal you have is your own past-guest data. I start with a lookalike audience (Google now calls these “audiences built from your data”) seeded off your strongest first-party list. In order of preference:
- Past direct bookers with a decent average daily rate. These are the guests you actually want more of.
- Newsletter subscribers who have opened in the last 6 months.
- Site visitors who hit your rooms or rates page but did not book.
Upload the customer list, let Google build a lookalike, and set the expansion to a middle setting at first. Then layer in interest and in-market audiences as a second campaign or ad group: “luxury travelers,” “in-market for hotels,” plus affinity segments that match your actual guest, foodies, wine enthusiasts, wellness, art and design, whatever your property genuinely is. Do not bolt on audiences that flatter you but do not match who books.
2. Geography: feed the markets your direct guests come from
Pull your booking data and find the top originating metros for your direct business. For a lot of independents this is a handful of drive-and-fly markets, not the whole country. I would rather spend a real budget saturating five feeder cities than a thin one nationwide. Exclude your own metro unless you have a genuine staycation or restaurant angle.
3. Creative: lifestyle, not a rate sheet
This is where most hotel campaigns die. People upload a logo, a room photo, and a “Book Now from $199” overlay, and wonder why nobody engages. Demand Gen is a discovery surface. Nobody on the YouTube home feed is shopping for a room. Sell the feeling of being there, not the transaction.
What I brief for:
- Short video, 6 to 15 seconds, vertical for Shorts and a square or landscape cut for the feed. Open on motion in the first second, the pool, the door swinging open, steam off the coffee.
- High-end stills of the experience: the room with the light right, the view, the bar, a plated dish. No price overlays.
- A soft line of copy that sells the destination, not the discount. “Your weekend in old-town” beats “Save 15 percent.”
- Multiple variants so the system has something to optimize between, and so you can refresh when one fatigues.
If your photography is dated, fix that before you spend a dollar on media. Good creative is the entire ballgame here, and it is also the cheapest lever. This connects to the broader content and reputation work that makes everything else perform.
4. Bidding and budget
Demand Gen has a learning phase. If you set the budget too low or change things every other day, you will keep it in learning forever and the results will look terrible. I run a steady daily budget for a few weeks before judging anything, usually starting on a clicks or conversions goal depending on whether I have enough booking signal to optimize toward. Patience is not optional with this campaign type.
The honest part: measuring view-through without lying to yourself
Now the hard bit, and the reason I see hotels either overspend on garbage or kill perfectly good campaigns.
Demand Gen’s whole value is upper-funnel. A huge share of its real impact is view-through: someone sees your gorgeous rooftop on YouTube on Tuesday, does not click, and three weeks later searches your name and books direct. Last-click attribution gives that booking entirely to your brand-search campaign and zero credit to the Demand Gen ad that planted the idea. If you only look at the Demand Gen line in a last-click report, it will look like a money pit.
So how do you measure it without kidding yourself the other direction (counting every coincidental booking as a Demand Gen win)?
| Signal | What it tells you | How much to trust it |
|---|---|---|
| View-through conversions | Bookings after an ad view, no click | Useful, but directional, not gospel |
| Branded search volume lift | Are more people Googling your name? | Strong, if the timing lines up |
| Direct booking share | Is the OTA mix improving over the window? | Strong over a long enough window |
| Geo holdout test | Causal lift vs. a matched control area | The gold standard, if you can run it |
| Last-click ROAS on Demand Gen | Almost nothing | Do not judge the campaign on this alone |
The two I lean on hardest:
Branded search lift. Pull your brand-name impressions and clicks from Search Console and your brand-search ad campaign. Chart them against your Demand Gen flight dates. If branded search climbs in your target metros while Demand Gen is live and settles when it pauses, that is real signal that you are creating demand. This is the cleanest proxy most independents can actually run.
Geo holdout. If you have the budget and the discipline, run Demand Gen in some feeder metros and deliberately not in matched control metros, then compare total direct bookings between them. Google also offers conversion-lift studies above a spend threshold. This is the only way to get something close to causal proof, and it is worth doing once so you know your true multiplier.
Rule I hold clients to: never evaluate a Demand Gen campaign on its own last-click revenue. Evaluate it on whether total branded search and total direct bookings went up while it ran. The channel that gets last-click credit is almost always brand search, which Demand Gen fed.
A realistic, clearly hypothetical example of how this plays out: say you run Demand Gen across four feeder cities for six weeks. The in-platform report shows a weak last-click return and a pile of view-through conversions you are not sure you believe. But your branded search clicks are up meaningfully in exactly those four cities, your direct booking share ticked up a couple of points, and your held-out control metro stayed flat. That is a working campaign, and the last-click number alone would have told you to shut it off. (Numbers here are illustrative, not a promise of what your property will see.)
Where this sits in the bigger picture
Demand Gen is not a standalone fix and I would never sell it as one. It is the top of a funnel that only pays off if the rest is built:
- It creates the branded searches that your local SEO and Google Business Profile work then captures.
- Those searches need to land on a site that converts to direct, which is the book-direct CRO job, otherwise you just fattened the OTAs.
- And the more your name and your story circulate, the more you show up in AI search and the LLM answers guests increasingly trust before they ever hit Google.
The realistic timeline: this is not a switch you flip for a booking spike next week. You are building name recognition that compounds. Expect to run a full learning cycle before you trust the numbers, refresh creative when it tires, and judge it over a quarter, not a weekend. There are no guaranteed outcomes here. What you can do is stack the odds: right audiences, distinctive creative, honest measurement, and a funnel that actually catches the demand instead of handing it to an OTA at 15 to 25 percent commission.
Done right, Demand Gen is how an independent hotel manufactures a little of the brand gravity that chains take for granted, and turns more cold strangers into people who one day search your name and book direct.
Want help building the top of the funnel?
If you want a Demand Gen build with creative that actually fits your property and a measurement plan that will not lie to you, that is exactly the kind of work we do. Book a free intro call and bring your booking data, we will sketch the feeder markets, the audiences, and the holdout test on the call.