Report

Long Tail effect on reservations

Hostel Trends and News

aboriginalhostel started this discussion 16 years 1 week ago

I was making some statistics about our reservation forecasts and surprisingly found this feature showing up in our numbers...

Basically you have an n number of reservations every month, which you have an x percentage of it are in month reservations.

According to my statistics, our reservations are:

n (total reservations) = n*x (in month reservations) +n*xˇ2 (next month reservations) +n*xˇ4 (reservations two months ahead) + etc.

The whole function shows a demand curve, which is exactly the same as long tail's.

Of course there are some small distorts in every month, but these are statistically neglected. However, if the curve drops instantly, that could be due to the economic crisis, explained in the Black Swan Theory. The crisis won't impact immediately but you can predict it's arrival from the changes of the fore coming reservations or you can check if you are already involved.

For example, our reservations curve is the following (no exact sales numbers are shown):

IMAGE(http://www.hostelmanagement.com/forum/attachment.php?attachmentid=195&stc=1&d=1226874405)

The months are related by percentage to each other.
As you can see, the light blue line at the front, which shows our last month reservations, 80% against 10%, the highest drop comparing to other months.

If I could only see that a month ago...

I was wondering if you have the time to check your own statistics and share your experiences on the forum... Maybe we could work out something about projecting reservations? ;)

Comments

Report

santa klaus

16 years

I have attached our "long tail" - the number of our beds booked until Feb 28.

It looks a bit less like a tail, but in principle, you are right.
It would be nice to compare that graph with the same one made exatly one year ago.

You´ll see that the tail gets very thin after January 5. I guess you wouldn´t be able to draw any meaningful conclusions by comparing it to the tail after January 5, 2007 - the numbers are just too small.
So - what you can do with it is looking 8 weeks into the future.

Anyway, even if I had that 8-weeks crystal ball, it would be hardly more than a toy for me. It would only make sense if I would still think about what price we should charge in December, but I have set our prices until the end of 2009 - we will only change them if something very, very dramatic happens. We have never had a situation dramatic enough to reconsider the price for the next month.

Report

aboriginalhostel

16 years

I have attached our "long tail" - the number of our beds booked until Feb 28.

It looks a bit less like a tail, but in principle, you are right.

Thanks for sharing, I knew I could count on you at least! ;)

However, the reservation forecast does not equal to your availability report.

Every month you receive a certain number of reservations. This would be your 100%. The majority of these reservations are for in month, which means the people who book will arrive the same month they booked. Our average is 60% of monthly bookings arrive in the same month they booked. About the rest, about 25% of the 100% on yearly average, reservations for the forthcoming month, 9% reserved two months ahead, 4% books three months ahead, 1% four months ahead, then again, around 1% each month ahead.

When you put these numbers (60, 25, 9, 1, 1, 1, 1, 1) into a curve, you'll get your tail. You can check the same feature by putting your own reservation numbers instead of percentages.

IMAGE(http://www.hostelmanagement.com/forum/attachment.php?attachmentid=198&stc=1&d=1227030818)

What I'd like to know, how can we predict by the forthcoming month reservations about how many guests will we have?

You´ll see that the tail gets very thin after January 5.

It would be nice to compare that graph with the same one made exatly one year ago.

You´ll see that the tail gets very thin after January 5.

As you can see, most of reservations (95%) are coming for the first three months, this is how far people book ahead. January 5th seems to be at the edge of this period, but you could compare your numbers with previous years statistics as well. If it was similar during the past couple of years, then there should be no problem.

IMHO this toy could help to check if your numbers are still okay.

It would be nice to compare that graph with the same one made exatly one year ago.

I did check those numbers and even if we had the same prices for the past 3 years, this years reservations was the worst.

Anyway, even if I had that 8-weeks crystal ball, it would be hardly more than a toy for me. It would only make sense if I would still think about what price we should charge in December, but I have set our prices until the end of 2009 - we will only change them if something very, very dramatic happens.

I think your most competitive town is Berlin, where your latter baby grows... Are you still successful over there, can you seal against the stream of lowering prices?
Because here we have about 80% of hostels having a price war against each other... but only 5 hostels have a full house so far. Unfortunately this year we have to be involved.

We have never had a situation dramatic enough to reconsider the price for the next month.

Knock on wood, I hope you'll be right!

Report

santa klaus

16 years

When you put these numbers (60, 25, 9, 1, 1, 1, 1, 1) into a curve, you'll get your tail. You can check the same feature by putting your own reservation numbers instead of percentages.

Ah, OK, I know what you mean. We once checked a large number of bookings and how far in advance they have been made and we came up with a tail like yours.
The problem is: our software doesn´t have reports like that. We had to do that manually. The data is there, buried somewhere in the databases, but there´s no tool to retrieve it, unfortunately.

What I'd like to know, how can we predict by the forthcoming month reservations about how many guests will we have?

Honestly, I think we can´t. Airlines and big hotel chains can do that, but not us. I remember I saw a documentary about Lufthansa: what struck me was that they have a team of 150 (!) persons doing nothing else than figuring out how much they can overbook each flight and how much to charge for the tickets. We don´t have that sort of manpower, we don´t have the proper software and we don´t have the heaps of data to feed it with. If we tried to do what they do, we´d only be awfully wrong too many times and make complete fools of ourselves. IMHO.

I think your most competitive town is Berlin, where your latter baby grows... Are you still successful over there, can you seal against the stream of lowering prices?
Because here we have about 80% of hostels having a price war against each other... but only 5 hostels have a full house so far. Unfortunately this year we have to be involved.

Well, we´re surely most competetive in Vienna - because there is so little competition here.
But! Berlin indeed comes in as close second and that´s a huge surprise.

I´m not exactly sure if we can seal against the stream of lowering prices, there are mixed signals. Our bookings on HW are 34% down compared to last month (as of today), compared to 19% down city-wide, but that could be a misleading figure as we are full already for new year´s eve while most other hostels are not. I assume many bookings made for Berlin now are for the last week in December.
On the other hand, occupancy is not bad at all there so far in November.

In any case, there is an unofficial club of Hostels in Berlin which apparently don´t care about the big price war - just like in Budapest. We aim to join that club, our rates will be kept the same throughout the winter and then we´ll see how we do.

Report

aboriginalhostel

16 years

The problem is: our software doesn´t have reports like that. We had to do that manually. The data is there, buried somewhere in the databases, but there´s no tool to retrieve it, unfortunately.

Indeed, having a software which supports custom reporting, should be priceless!

Another advantage for us, being a small hostel: administration work is much smaller! :D

However, I did the job myself manually, within couple of hours, including the past few years, although we have about 1000 reservations a year excluding walk ins and e-mail bookings.

What I'd do in your case instead of doing the whole thing manually, is exporting your bookings into a monthly breakdown excel sheet from your HW inbox, and asking your geek to figure out how to create functions in order to get your reservation forecast. IMHO that would work.

HW is a bit inappropriate, as less people book online during off season but still a good indicator.

Honestly, I think we can´t. Airlines and big hotel chains can do that, but not us.

I thought you are selling 35,000 bednights a month. IMHO, that's an ideal large number for statistics, much better than out 800 bednights.
Large numbers never lie!;)

I remember I saw a documentary about Lufthansa: what struck me was that they have a team of 150 (!) persons doing nothing else than figuring out how much they can overbook each flight and how much to charge for the tickets.

Well, my version is rather simple. We start with previous year prices, just like you. If needed, I check the prices on HW (100+ is a large number of hostels as well) and aim for the upper 70% region. This is how much our reputation, marketing and the amount of visitors could combine up to sell our beds. Pretty X factors all but guessing is a great invention. :)
This year we had to go down to 50%, due the lack of visitors and the tougher price war. Two weeks ago we had no bookings ahead. Today we have a full house. Predicted sales are -10% comparing to last year but at least it's still +20% comparing what we couldn't make without these guests.

We don´t have that sort of manpower

Neither us, but we both have computers which can calculate much better than both of us. The question is always how to feed them with data. I think your geek can help with this but if not, and you think it is still important, you can still ask someone to help you with! ;)

We have a quote on the reception door: We can provide you the impossible but you may have to wait for miracles! :)

we don´t have the proper software and we don´t have the heaps of data to feed it with. If we tried to do what they do, we´d only be awfully wrong too many times and make complete fools of ourselves. IMHO.

Well, you don't need to do the same what they do. We are still far from their pricing but the main goal is to sell the most of beds with the most of revenue.
As a simple equation, 100% capacity * 80% price equals 80% capacity * 100% price. Anything above is a profit.

As long as you can sell the beds you want, there is no need to worry and if not, then there will be always channels how to sell'em. This is how the market evolves.

I´m not exactly sure if we can seal against the stream of lowering prices, there are mixed signals. Our bookings on HW are 34% down compared to last month (as of today), compared to 19% down city-wide, but that could be a misleading figure as we are full already for new year´s eve while most other hostels are not. I assume many bookings made for Berlin now are for the last week in December.
On the other hand, occupancy is not bad at all there so far in November.

The same thing I thought in October. We had a good month as usual. The only problem I haven't seen until recently, even we had a slight change on number of reservations made during October comparing to last year, reservations for November were halved.

An availability report could show a parallel curve with number of visitors when the competition have static prices (as mostly with few competition), but in such competing city as ours, I think this toy is more reliable, IMHO.

Report

santa klaus

15 years

What I'd do in your case instead of doing the whole thing manually, is exporting your bookings into a monthly breakdown excel sheet from your HW inbox, and asking your geek to figure out how to create functions in order to get your reservation forecast. IMHO that would work.

HW is a bit inappropriate, as less people book online during off season but still a good indicator.

Yes, it may be inappropriate. The HW inbox has a lot more reports that our reception software, so I´m tempted to use those HW reports. We only sell half of our beds through HW though, so what I see there ignores the other half. Furthermore, if HW sales change, there is no way to say if that´s a HW problem, a Wombats problem or a general problem.

I thought you are selling 35,000 bednights a month. IMHO, that's an ideal large number for statistics, much better than out 800 bednights.
Large numbers never lie!;)

OK, 35,000 is better than 800 - especially since these 35,000 are in three destinations instead of just one. I would be much better off statistic-wise with 800 each in 44 destinations.
In addition to that, it´s nice to compare those 35,000 with what was going on in the same months in previous years. For Berlin, we have nothing at all to compare them with, in Munich only 4 years and in Vienna it´s even more complicated: we started with 160 beds in 1999, expanded to 260 in 2003 and added a second house in 2006. We started with taking lots of groups in 1999, then reduced them as more backpackers came, then increased groups again when we added 3 floors only to see that now even more backpackers came, just to reduce groups again, then the same thing all over again when we added the second house. It´s sort of worthless to even compare June 2006 with June 2007 as in the latter we had less groups and more backpackers.
Large numbers per se don´t lie but you can prove everything or nothing with them.

Well, my version is rather simple. We start with previous year prices, just like you. If needed, I check the prices on HW (100+ is a large number of hostels as well) and aim for the upper 70% region. This is how much our reputation, marketing and the amount of visitors could combine up to sell our beds. Pretty X factors all but guessing is a great invention. :)

My point. It´s nice to have large quantities of numbers to crunch, but there´s the danger that you sometimes don´t see the forest because of all the trees. A good guess can often be better.

An availability report could show a parallel curve with number of visitors when the competition have static prices (as mostly with few competition), but in such competing city as ours, I think this toy is more reliable, IMHO.

I don´t know. Berlin has at least as much competition as Budapest. We have static pricing and quite high prices now as well, yet we do extremely well, even now in Winter. I refuse to get obsessed with pricing, there MUST be more important factors.
What I have on my to-do-list is to molest our guests when I see them booking accomodation on our internet terminals. I want to ask at least 50 just how they make their decisions where to stay.

Report

aboriginalhostel

15 years

Yes, it may be inappropriate. The HW inbox has a lot more reports that our reception software, so I´m tempted to use those HW reports. We only sell half of our beds through HW though, so what I see there ignores the other half.

So, when I said maybe it is time to replace your front desk software, this is what I meant... ;)

Furthermore, if HW sales change, there is no way to say if that´s a HW problem, a Wombats problem or a general problem.

As we both know, this part requires complex and dimensional thinking, which is hard to model and extract into any other test environment. So guessing is about doing the same thing intuitively. It's like thinking ahead few steps in chess, counting variations as much as possible and trying to push the game to the preferred direction, however there is still one result of a game, regardless how many variations could be played.

I like theories and also to confute them right away: It is hard to state something certain but gives clues to follow...

If sales are down, it could be depending on less tourists in budget sector, the economy crisis, budget flight companies failure, BE's failure, Budapest's failure, better competition, price wars, currency rate changes, Aboriginal hostel's failure, recent bad reviews, etc. Complex enough to find one reason through these variables, but getting through each of them and searching feedbacks could help a lot.

I just came back from Brussels and met couple of European hostels talking about how sales go down everywhere. So it is not Budapest only then. I'm happy to see that despite our sales dropped down at the beginning of the month, by the end we will have the same revenue as last year.

OK, 35,000 is better than 800 - especially since these 35,000 are in three destinations instead of just one. I would be much better off statistic-wise with 800 each in 44 destinations.

I agree, that's why I shared this info on the forum. I can also state that December will be worse comparing to November (except New Years Eve of course) but January will be much better comparing to last year.

In addition to that, it´s nice to compare those 35,000 with what was going on in the same months in previous years. For Berlin, we have nothing at all to compare them with, in Munich only 4 years and in Vienna it´s even more complicated: we started with 160 beds in 1999, expanded to 260 in 2003 and added a second house in 2006. We started with taking lots of groups in 1999, then reduced them as more backpackers came, then increased groups again when we added 3 floors only to see that now even more backpackers came, just to reduce groups again, then the same thing all over again when we added the second house. It´s sort of worthless to even compare June 2006 with June 2007 as in the latter we had less groups and more backpackers.
Large numbers per se don´t lie but you can prove everything or nothing with them.

Indeed it is hard to find the same patterns in an ever changing environment, even if it's about your own hostel. What I did about ours, comparing just the same rooms in a monthly breakdown for the past two years. We had exactly the same numbers going on for the summer, and even some winter months got up 25% comparing to last year. Overall 15% better than last year, but the same as last year due we are less an apartment this year. Of course it is not the same sales as the very first year as we are competely changed, but at least I can compare the last two years...

My point. It´s nice to have large quantities of numbers to crunch, but there´s the danger that you sometimes don´t see the forest because of all the trees. A good guess can often be better.

IMHO guessing is not just random but an intuitive thinking from the back of your mind. You can guess more precisely if you have more, relevant information. IMHO.

I don´t know. Berlin has at least as much competition as Budapest. We have static pricing and quite high prices now as well, yet we do extremely well, even now in Winter. I refuse to get obsessed with pricing, there MUST be more important factors.

That's interesting... I met an Irish guy on the plane home, talking about hostels in general. He said as long as you set up a brand you can charge the same because you get a certain standard everywhere you go, like Hilton and Marriott works. It is definitely an advantage for the chains.

What I have on my to-do-list is to molest our guests when I see them booking accomodation on our internet terminals. I want to ask at least 50 just how they make their decisions where to stay.

Sounds plan... Let us know the results if possible! I'll do the same thing and we can compare them!

Search form

Report 'Long Tail effect on reservations'?

Report 'Re: Long Tail effect on reservations'?

Report 'Re: Long Tail effect on reservations'?

Report 'Re: Long Tail effect on reservations'?

Report 'Re: Long Tail effect on reservations'?

Report 'Re: Long Tail effect on reservations'?

Report 'Re: Long Tail effect on reservations'?

Join the world's largest community of hostel professionals.

Check Your Email