September starts in six days.

Two months ago, in The August Illusion, we argued that peak season hides your real marketing performance — that demand is loud enough to mask whether your brand is doing any work at all, and that the truth only surfaces once the noise clears.

This is the edition where the noise clears.

First: you probably didn't track it

In Issue #32 we asked for thirty seconds a morning. Two numbers. Total bookings yesterday, direct bookings yesterday.

Most of you didn't do it.

That's not a criticism. It was August, and August does not leave room for new habits.

But you can recover almost all of it, and you have six days.

Open your PMS. Export August bookings with booking date and source. Group by day.

You now have the same 31 data points, retroactively.

It takes about twenty minutes. Do it before September 1st — because the moment demand softens you will start making decisions, and you want a baseline to make them against.

The three patterns

Once the daily series is in front of you, one of three shapes will emerge.

Pattern A — direct share holds steady across the month.

Travellers sought you out regardless of how tight availability was. This is the strongest signal a hotel can get: your brand did real work during the one month when it didn't have to.

Pattern B — direct share is high early, then falls away.

In the first half of August availability was scarce and travellers stopped comparing. In the second half supply loosened, they went back to comparison shopping, and you weren't the property they remembered.

Your direct share was riding market pressure, not brand pull.

Pattern C — direct share is low throughout.

Your rooms filled, but almost none of it came from your own work. You were found, not chosen.

September will confirm this, and it will confirm it expensively.

Most hotels in Greece will end August somewhere within these three patterns.

Very few will know which one.

The month that reveals your marketing performance is not the month you can fix it in.

Now the distortion

Here is where most diagnoses go wrong.

There is a widely held assumption in hospitality that "direct" means "free" — that a booking arriving through your own website carries no acquisition cost, and therefore every direct booking is a win.

It isn't true.

Direct bookings carry real costs. The booking engine. The website. Google Ads. Metasearch bidding. SEO. Payment processing. The agency retainer.

A hotel bidding aggressively on its own brand name is paying to acquire a guest who was already coming. That booking lands in your system as direct.

It is not brand pull. It is paid acquisition that reports as direct.

Which means the number you are about to diagnose may be inflated.

If your August direct share was 22%, and eight of those percentage points arrived through paid search and metasearch, your unpaid direct share was 14%.

Those are two very different hotels.

The stricter number

This is why Issue #32 asked specifically for bookings that arrived without an OTA and without a paid campaign.

Not because paid acquisition is bad. It's legitimate, often necessary, and frequently the right decision.

But it answers a different question.

Paid direct tells you how efficiently you can buy demand.

Unpaid direct tells you how much demand exists without you buying it.

The second one is the cleaner signal of organic demand and brand pull — demand that exists because the traveller wants you, not because someone paid to put you in front of them.

It's the one that compounds.

It's the one that still works in a month when you pause the ad spend.

So when you build your August series, split it into three lines:

The gap between the second and third line is the closest thing your August data gives you to a read on brand pull.

A word on the mix

There is a fair argument, made well by people who work in distribution, that OTA, B2B and direct aren't competing channels at all. Each does a different job. OTAs acquire guests and deliver visibility at scale. B2B opens markets a property can't reach alone. Direct owns the relationship and the data.

That's correct, and Destinova has never argued for zero OTA.

The argument here is narrower: most Greek hotels are structurally over-indexed on acquisition and under-invested in ownership.

The mix isn't balanced. It's tilted — and the tilt is expensive.

The September diagnosis doesn't tell you to abandon channels.

It tells you where your tilt currently sits.

What each pattern asks of you in September

If you are Pattern A — protect it.

The instinct in a softening month is to discount, and discounting is the fastest way to teach a loyal audience that your rate is negotiable. Hold price. Keep communicating. Your brand is working; don't undermine it in the first slow week.

If you are Pattern B — the work is positioning and memory.

Travellers found you when they had no better option and forgot you when they did. That is not a pricing problem or a service problem. It's a recall problem — travellers didn't think of you when the moment came to choose. And recall is built in the months when travellers have time to think. Which begins now.

If you are Pattern C — this is a twelve-month project, not a September campaign.

You are structurally dependent on other people's distribution. The fix is building an audience you can reach directly, and it starts in October, not next May.

The uncomfortable part

None of those three answers is a September tactic.

That is the real content of the diagnosis.

September tells you what happened. October through April is where you do something about it.

Which is why the hotels that grow year over year are so quiet in August — they did the work eleven months earlier, and August is simply when the work shows up as revenue.

August fills your rooms. September tells you why. And the answer was decided long before either month began.

💬 Worth asking yourself this week

When you subtract every booking that arrived through an OTA, and every booking that needed a paid click to happen — what's left?

That number is your clearest signal of brand pull. Everything else is distribution or acquisition.