[MUSIC]

In this video we will be discussing where to get data and why it is critical to

put this data into context which helps turn it into information.

Let's start with where to find the data.

My friend Anton Helf from Hit Consult was

kind enough to let us use this fantastic picture.

This diagram shows the main technologies used by a small

to medium sized hotel chain in Europe and is based on an actual chain.

The first thing that you might notice, is the hotel industries never ending love for

three letter abbreviations goes much further than key performance indicators,

or KPI's, and extends all the way To the systems we use.

The second thing you might want to know is that although there are lots of different

arrows showing some kind of interface, or connection between the different systems,

each system is really its own silo or islet.

And there is really important information in many of these systems.

Actually, it's my job at SnapShot to work with each of these data silos

to get that valuable information into our system so

that hotel managers can make right data-driven decisions.

Without having to dig into each of these silos individually.

Finally, you might notice that the most connected system is probably

the property management system and in fact for most hotels the most important

data source is their very own property management system.

In general, all reservations flow into the PMS, and this is the first building block.

Now a number of fields on a reservation that must be imputed manually or

over an interface.

These will help us do the analyses we need later.

In well managed hotels, each reservation is actually checked by

each shift to make sure that it's complete and has the correct information.

The rate code is connected to a specific price.

For example, a negotiated seaman's rate of 80 euros.

This could also be a flexible rate, such as a BAR or best available rate.

Market segments allow you to categorize the guest into different types.

Typical segments include corporate negotiated or

uncontracted transient lesson.

There is usually a connection between the right code and the market segment.

So if your PMS allows it, it's usually a good idea to have the market segment

automatically selected on the basis of the rate.

As we saw in MOOC 2 on the fundamentals of revenue management,

since different segments show very different booking behavior,

it is extremely important to keep the segments separated.

For example corporate guest with negotiated rates typically book

much more last minute than non-negotiated transient individuals.

At an absolute minimum,

group reservations must be held in a different segment than individuals.

It also makes sense to separate any negotiated ranks

from publicly available rates.

The source field is there to allow who tells to track which

channel the reservation came from.

Many hotels struggled with this one

because they are still using source code set up from years ago.

And use buckets like guest direct and online which failed to tell you for

example that the reservation came from booking.com.

Now that there has been an explosion of channels and

especially with the differences of cost distribution between different channels,

it has become even more important to add the granularity to the source.

Most PMSes do allow you to create a profile.

For example, for each agent or agency, like booking.com.

Some hotels do use this to track how many reservations have been made

by that online travel agent.

However, in my experience it's actually not a good idea to rely on these agent

profiles, since not all reservations need an agent attached, and it's really

easy to forget to add an agent profile which really destroys your data quality.

Typical reservation types are guaranteed by credit card or 6 PM arrival.

These are important to track, as they will have very different cancellation behavior.

Guests that cancel late or no-show and

have a guaranteed room can be charged for the no-show.

6 PM arrival reservations means that the hotel

only guarantees to hold the room until 6 PM.

At which point if the guest has not arrived you can try to resale the room.

One of the most important comparisons for a revenue manager is to know how many

reservations we were holding at the same time last year for the same arrival date.

Well, actually, not for the exact same arrival date, but

adjusted by one day to account for the shift in the day of the week.

It really makes no sense to compare last year's Sunday with this year's Monday,

due to significant differences in the day of week pattern.

You can see on this example that to get Tuesdays, the Tuesdays to match up,

we need to compare the second of September to the first of September.

Unfortunately, most property management systems do not store this information.

Which means, that if you would like to do this comparison you will have

to store the information yourself.

This means that each and every day, also on the weekend,

you need to export the number of reservations on the books for

every day into the future where you're accepting reservations.

This data gets pretty big pretty quickly.

For a single year, you would be storing 133,225 numbers.

That's 365 times 365 without even a break down by market segment or channel.

If you only had five segments and five channels then we're

talking about over 1.3 millions numbers you generate as

every year just to get a single snap shot of the situation once a day.

It is really important to export this data at the same time every day.

Since the reservation situation changes constantly over time.

And you really wanna compare apples to apples as much as possible.

At the beginning of this session you saw a diagram of many different systems that

provide valuable information but they are all stuck in silos.

It is critical that these different sources are brought together at a single

location for a number of reasons.

Perhaps the most important reason though is simply how our brains work.

We actually aren't really good at saving information in our brains,

especially when we need to compare one screen to the next.

Let's play a game to prove this to you.

You will see a picture flashing on the screen.

Try to identify what has changed in this picture

It's pretty difficult to see right?

The first time I saw this I couldn't find it.

Now let me make it easier for you.

Now you see it right away right?

Without the quarter of a second of black in between it is so obvious and

easy to see.

So even if just a split second of black is enough to confuse us,

how in the world are we going to be able to look at a report from one system and

connect the dots to another system.

For human understanding we need to aggregate our data into one view

to allow us to really understand it.

Context is king.

In the next video we will take our first look at analysis.

In this case benchmarking of performance between similar hotels.

Data Collection and Aggregation

Demand management: Breaking down today’s commercial silos

Meet the Instructors