4. THE HIVE, IN-BRIEF
1
2
3
Hostel that provides LOW-COST accommodation
to backpackers and travelers around the world
Total of 15 rooms
Best accommodation, Lowest prices
12. Using an Inventory Planning System,
we minimise wastage, reduce stockouts.
0% FOOD WASTAGE
With an Inventory Planning System,
we minimise wastage, reduce stockouts.
Level of Inventory that triggers an order
for additional stock
19. Problem
Check-in Date Name Amount Paid (S$)
Mon Ah Kow 45
Tues Barnabas 45
Wed Ma Lian 90
Thurs Maki - San 135
Fri Elon 90
$135, Receipt to Ms Chick
$45, paid on Wed
$90, billed on Sat
Manager scours through
daily records to tally with
cash/check records
Automating Guest Check-ins
20. Check-
in Date
Name Nights
Amount
Paid (S$)
Mode of
Payment
Payment Settled?
Mon Ah Kow 1 45 Cash Y
Tues Barnabas 1 45 Master Y
Wed Ma Lian 2 90 Visa N
Thur
s
Maki - San 3 135 Cash N
Fri Elon 2 90 Check Y
Solution
A spreadsheet that correctly
captures the payment made
by every guest in real-time
Automating Guest Check-ins
30. Using historical data, we construct a trend analysis.
Exponential Smoothing
With Linear Trend
F(t) = αA(t)+(1−α)[F(t−1)+T(t−1)];
T(t) = β[F(t)−F(t−1)]+(1−β)T(t−1);
f(t+τ) = F(t)+τT(t), τ =1,2,...;
33. Optimal Room Pricing
No framework in organizing prices!!!
Reliance on intuition and gut feel
Might be under-charge during peak seasons
Afraid to raise prices excessively
1
2
3
4
34. Optimal Room Pricing
Hostel under-going expansion in
converting all rooms to dorms
Wants a price that maximises price and
capacity
35. Optimal Room Pricing MODEL
–Rela&onship
between
price
and
occupancy
–Cost
data
associated
with
occupancy
changes
Pricing framework that optimizes profits
Maximizing total revenue relative to total cost
• Need
to
es&mate
36. 3
How do we find the
OPTIMAL PRICE?
Relationship between price and
occupancy
Cost data associated with occupancy
changes
38. Historical Panel
Data on price and
occupancy
Perform Linear
Regression
Tease out
relationship
between price and
occupancy
Optimal Room Pricing MODEL
Estimating Demand:
39. Multi-Variable Regression
Priceit = β0 + βoccOccit + βfebFebi + βmarMari + βaprApri + βmayMayi
+βjunJuni +βjulJuli + βaugAugi + βsepSepi + βoctOcti + βnovNovi +
βdecDeci + βincincit + βprice
2 Price2
it + uit
Regression Model:
OBSERVATIONS
• t-stat for βprice
2 & βinc insignificant at 5%.
–Drop Incit & Price2
it from the model
–Linear demand curve is justified
60
40. Priceit = β0 + βoccOccit + βfebFebi + βmarMari + βaprApri + βmayMayi
+βjunJuni +βjulJuli + βaugAugi + βsepSepi + βoctOcti + βnovNovi +
βdecDeci + βincincit + βprice
2 Price2
it + uit
• F-stat for month variables are significant at 1%
– Cannot drop Febi … Deci from the model
– Able to observe monthly demand curve
Multi-Variable Regression
Regression Model:
OBSERVATIONS
60
41. Data: Variable Cost
• Breakfast
• Need
to
es&mate
marginal
cost
pertaining
to
breakfast
per
addi&onal
consumer
• Issue:
–Backpackers’
appe&te
vary
–Hence
consump&on
paCerns
vary
from
month
to
month
42. • Monthly
breakfast
data
in
2012
• Proxy
for
future
monthly
breakfast
consump&on
paCerns
Data: Variable Cost
• Constant
Unit
Cost
(per
guest)
–Laundry
• Assump&on:
–Same
contractor
will
be
engaged
for
the
foreseeable
future
43. • Cost
factors
that
do
not
vary
with
occupancy
–Bed
frames
–Pain&ng
–Ligh&ng
–Air-‐condi&oning
• Assump&on
for
monthly
fixed
cost
–Straight
line
deprecia&on
Data: Fixed Cost
44. Setting up the model: Cost
Input:
• Cost:
• Breakfast items
• Utilities
• Estimates:
• Useful life
• Disposal value
• Month
• GST
• Service Charge
Intermediate Outputs
Variable Cost per head
Fixed Cost
Levy per head
45. Input:
• Demand function
• Room rate
• Month
Constraints:
• Max occupancy
Intermediate Outputs:
Total Revenue
Occupancy
Setting up the model: Revenue
47. Constrained Optimization
• Use of solver to find:
–A room rate that maximizes total profits
–Subjected to maximum occupancy constraints
• Performance variables:
–Profit
–Room rate
• Consequence Variables:
–Total revenue
–Total Cost
–Occupancy
49. • Expansion plans
–Constraint value
–Input more fixed cost items purchased
• Inflation/ Change of suppliers
–Breakfast cost items
Model Flexibility
50. • Monthly analysis
–Month input from drop-down list
• Change in policies/seasonal levies
–GST input
–Service Charge input
Model Flexibility
51. Lack of industrial data &
knowledge.
Difficulty in conceptualising
relevant variables and
integrating models.
LEARNING JOURNEY
In the end, we learn how
powerful and beneficial a
simple program like Excel is.
53. BUT WE WOULD REALLY LOVE
TO HEAR WHAT
YOU*
HAVE TO SAY ABOUT IT.
54.
55. Assumptions
• 1. Lead time is constant
• 2. Inventory carrying cost per unit of item
do not vary
• 3. Monthly consumption patterns should
be similar to previous years
• 4. Variability of consumption each month
is similar
56. • Hostel relies heavily on a high no. of
regulars
–Come at specific time periods of the year
–Same backpacker’s appetite do not change
• Future breakfast consumption will follow
similar patterns
• Reasonable assumption
• Most reliable estimate
Assumptions
57. Further Assumptions
• Manager to run model only at the start of
every month
• Other independent costs are excluded
from the analysis
–Requested by owner
58. Model Limitations
Reorder Point
The reorder point model can only be used in cases where ordering costs, lead
time and demand are constant.
Occupancy
Forecast
Despite using Exponential Smoothing to forecast the trend of the occupancy
rates, there is always uncertainty involved in predicting occupancy rates. There
are many other factors involved such as economic conditions of the tourism
market, affluency rates and presence of competition. Hence, the model is used
only for estimation purposes and should be treated as such.
Optimal Bed
Pricing
This model can only be run at the beginning of the month. It will not export
accurate results if it were to be run at any other point during the month.
Visual Basic for
Applications
(Automating Guest
Check Ins)
The limitations of Excel as a database management system (DBMS) are quite
apparent in this project. While it is effective in handling raw entry and tabulation
of guest data, it is inflexible in allowing the user to edit information that has
already been entered. This modification anomaly commonly present in many
DBMS cannot be resolved by the Excel model alone.
Also, as the system does not support client-side validation for reservations,
erroneous entries made by the staff might be picked up, leading to inaccuracy. In
the long run, the lack of accurate data input or consistent updates may lead to
compounding inaccurate trends.
Looking forward, as the Hive expands, it may consider more complex DBMS
solutions such as SQL and Oracle to give the owner of Hive greater flexibility in
managing the hostel’s guest data.