final Flashcards
services, operations, service operations
services: stuff we buy that we don’t get physical stuff in return; if customer is not present, something belongin to the customer is present
operations: business processes involved with creating/delivering/providing a product and/or a service
service operations: business processes involved with creating/delivering/providing a service
3 sectors of the economy
how did this change over time?
primary— agriculture, fishing, mining; extraction of raw materials
secondary — manufacturing, construction; processing of raw materials/semi-finished goods
tertiary — services
over time, went from primary (agrarian society), secondary (industrial revolution) to tertiary (service focused)
characteristics of services
simultaneity — you are receiving/consuming the service at the same time it is being provided to you
* ex. haircut
* ops challenges: slow times between meals
perishability— services perish instantly because they are time-based
* if you are not providing a service, you can never get that time/capacity back
* ex. empty airplane seats
intangibility —nonphysical
* ops implications: hard to convey the quality
OM triangle for services
capacity
information
queues (not inventory!)
metrics
metric, utilization
are OEE and TEEP good for services?
metric: measure of an aspect of a business process; usually related to strategic, tactical, or operational goals
utilization: amount made / total amount that could be made; can go over 100% in services
no bc demand fluctuates
metics
OEE + how it can be manipulated
how to make it better? what does speeding up processes do?
OEE: overall equipment effectivenesss; availability * performance * quality = A * P * Q
availability= operating time (actual) / scheduled time (predicted)
performance = theoretical time (units made / cap rate) / operating time
quality = good units * total units
failing to add scheduled overtime makes A look better, making OEE bigger
add overtime to make A better; speed up makes op time decr but more defe
metrics
TEEP
adv of TEEP over OEE
total effective equipment performance
TEEP = loading * OEE
loading = scheduled time / calendar time
calendar time = 24 * 7 * 60
harder to manipulate bc scheduled time is canceled out
metrics
profit-per-partner
= margin * productivity * leverage
margin = profit / fees (revenue)
productivity = fees / staff (excl partners)
leverage = staff / partners
= profit / partners
process analysis
flowcharting: 🔵 ⬜️ 🔷 🔻 →
swim lane flowchart?
🔵 start/end
⬜️ operation
🔷 decision
🔻queue/buffer
→ flow
swim lane: columns represent departments/employees/resources
why do we prefer a smaller buffer?
link to JIT
- might have a space constraint
- smaller buffers decrease total time
- less mistakes to correct within a single smaller buffer
- why manufacturing companies move towards JIT
process analysis
job shop
service examples, challenges
custom orders
- different requirements and different paths through the system
- a lot of flexibility, variety
- resources organized by specialty/function
- ex. surgery
- challenges: variability in supply, balancing resource utilization, guiding customers through process
process analysis
project, continuous flow
project: individual, one-time; one of a kind
* challenges: scheduling deadlines, assigning resources
* ex. IT project, construction
continuous flow: not individual units; uninterrupted delivery
* ex. internet, cell data
* challenges: capacity planning
process analysis
batch flow
things done in batches; served in groups
* less customization
* not as efficient as assembly line
* ex. rollercoaster, movies, transportation
* challenges: pricing per person, creating the batches
process analysis
assembly line
identical products + processes; cutsomers follow the same sequence
* challenges: balancing resource utilization along the line, meeting demand during peaks, not much flexibility
* ex. fast food restaurants, car wash
what questions to consider when collecting data to create a forecast? what to consider when doing forecasting?
data collected:
* how specific/detailed/aggregated?
* quantitative, qualitative
* time scales -> ordering cycles, staffing, hourly, weekly, monthly, yearly
things to consider:
* what it will be used for
* other info needed
* info you can ignore
processes that generate demand
mkt forces, trends, weather/external factors, competitors’ actions, price changes, illness (in health fields)
explain the challenge of censored demand
retail example, transit example
past data often has a cut off, pieces missing
data that wasn’t collected/wasn’t satisfied
retail: what people wanted but couldn’t find isn’t recorded
transit: people who don’t get on the bus isn’t recorded
EOQ assumptions, when we use EOQ, insights
assumptions: stable, predictable demand
use when: relatively flat demand, fairly long shelf life
insights: tradeoff between FC and VC
newsvendor assumptions, when to use, insights
assumptions: short shelf life, variability in demand w a known distribution
when to use: perishable products, or when we have to decide how much to order and can’t make any adjustments to that cycle
insights: trade-off between ordering too little and too much
forecasting and inventory: things to consider
- how do these activities affect each other? -> forecast is an input into inventory planning
- what if the forecast is wrong?
- substitutable products complicates inventory
ezza: operations challenges
- variability: chipping
- variability: seasonality of demand
- expand to pedicures? > changes their model/reputation
- staffing > takes a long time to become a nail tech
ways to adjust supply: short-term capacity changes
healthcare, sushi, hotel examples
- add production time
- remove producton time
- shift capacity from/to other products
- outsourcing
outsource; more empl/move tables; change room types/staff/outsource
adjusting demand
healthcare, sushi, hotel examples
managing demand: changing what demand is; affect demand using promotion, advertising, cross-selling
* ex. happy hours
planning for unsatisfied demand (choosing to not be able to meet all demand); can choose who you satisfy
prioritize severe injuries, happy hours, promotions/dynamic pricing
asynchronicity
healthcare, sushi, hotel examples
inventory: supply occurs before demand; made in advance
backorders: demand occurs before supply, ex. pre-orders
* services: queues
hospital scheduling/waitlists; reservations; vouchers
JIT, jidoka, pull system
kaizen, heijunka
JIT: make only what is needed, only how much is needed, only when it is needed
jidoka — build quality into the process; anything that doesn’t add value is waste
* allow problems to be solved immediately by pulling andon cord
* stop the system, fix mistakes, then continue
* use visual systems to monitor/control processes throughout
* inspections at end of process = wasteful
* 5 why’s → takes longer but addresses the root issue
pull system: very little inventory at each station
kaizen: continuous improvement through job satisfaction (rotation)
heijunka: spread out demand as evenly as possible; all the different models made at the same time
dabbawala: how do they achieve such a high level of service? what types of redundancies/variability exist?
service level: workplace culture (religious), committed workforce (not many opportunities elsewhere), cross-training, extra workers at the platforms, colour coding on the dabbas
redundancies: double sorting, backup workers
variability: weather, road and train repairs, bike repairs, delays in pickup, customers!!
job shop layout: goals, strategies
when do we use it? in services? other solution approaches?
goals: to minimize walking, distance the product travels
* works for things like zoos where people can choose their path; not for things based on steps like subway
* services: how to arrange the individual services within a service providing-organization, relative to each other
strategies: move one unit, swap 2 units
rectilinear/taxicab distance — go along the grid
1. add traffic between each area -> upper right triangle
2. figure out the distance between each combo within the proposed layout and add them to the matrix triangle
3. SUMPRODUCT the totals with the distances for the whole upper triangle
4. optimize by minimizing the total traffic
incr improvements, simulations, IP
problems with TTC
what does it mean when the customer is in the factory?
- provide bus service to low-density routes > low utilization
- lack of customer service policies > exacerbated by social media
- escalating tensions between union and company > uncivil interactions
- lack of funding leads to lower service levels, more delays, leading to even more complaints
- fare increase
customers watching the service being performed
how do we measure quality?
hospitals?
- on-time service
- service level (not stocking out)
- customer perceptions of quality: ambiance, friendliness, timeliness, value, long-lasting
- hospitals: readmissions, wait times, health outcomes, patients’ family members, employee satisfaction
4 things to consider in process improvement
flow, resources, queues, capacities
foundations of process improvement
- customer satisfaction — focus on customer needs
- management by facts — use data, scientific thinking, statistical analysis
- respect for people — improvements occur at, and with support from, all levels of the org
six sigma background
DMAIC, effectiveness equation, hawthorne effect
- comes from manufacturing → quality control background
- focused on process improvement + reducing variability
- remove the causes of defects/errors
- relies on all levels of management → strong leadership is key
- sigma = measure of variation (SD); want to scrunch the dist together to reduce variability
DMAIC — define, measure, analyze, improve, control
Q * A = E > effectiveness needs quality and acceptance > culture change crucial
indiv modify their behaviour bc they know they’re being observed
six sigma: CAP
change acceleration processes; ex. elevator pitches, process mapping, communications plans, stakeholder/resistance analysis
lean (process improvement) background
- any resources that are being spent on activities that don’t directly add value are wasteful
- derived from TPS
- identification and steady elimination of waste
- often applied to business processes in services
- team/group initiatives → involving everyone who touches a process to figure out ways to improve it
pareto analysis background
what does the line represent
bar chart for count for each category of complaints, sorted highest to lowest
* cumulative values are plotted with a line
SERVQUAL background
what scale is used?
22-question survey
intended to work across various companies and industries
can add more questions based on industry
change how the questions are asked
uses the 5 service quality dimensions as broad categories
rating on a Likert scale
type I errors, type II errors
whose risk? what is statistical process control?
type I errors: identify process as out of control when it is actually OK
* false positive, producer’s risk
type II errors: define process as OK when it is actually out of control
* false negative; consumer’s risk
identifies when a measure is out of control; metric w repeated values
R-chart vs X-chart
where is the horizontal line? LCL? UCL?
R-chart: measure range over time
* R̄ at avg range
* LCL = D3R̄
* UCL = D4R̄
X-chart: measure average x for each subgroup
* X-barbar at avg avg x
* LCL = X̄̄ - A2R̄
UCL = X̄̄ + A2R̄
data envelopment analysis: what does it do, what is efficiency
relatively efficient vs inefficient units
DEA compares the efficiency of multiple service units that provide similar services by explicitly considering their use of multiple inputs to produce multiple outputs
* efficiency is a score/grade defined by the input/output ratio
* incorporates multiple inputs and multiple outputs into both the numerator and the denominator of the efficiency ratio
* compares a particular unit’s efficiency with the performance of a group of similar service units that are delivering the same service
relatively efficient units = 100% efficiency
* not allowed to use a formula where one unit gets > 100% efficiency
inefficient units < 100% efficiency
what is the DEA productivity frontier? where do we want to be?
shows combinations of inputs
want closest to the origin → uses less inputs
different shapes based on what we are measuring
efficiency reference set, composite reference unit C
efficiency reference set — relative weight assigned to that efficient unit in calculating the efficiency rating
* shadow prices associated with the respective efficient-unit constraints in the solution
composite reference unit C — defined by the weighted inputs of the reference set; defined at the frontier and the inefficient unit
approaches to facility location
cross-median, competitors, convenience
cross-median approach: how it works, assumptions, limitations
find a line such that >1/2 are on either side
1. plot population points
2. add up population and divide in 2 for median
3. add up the populations up to at least the median starting from both sides of the x and y axes (can tilt axes)
4. get 2-4 medians
assumptions: population is in the middle, evenly distributed OR that reaching the first point = reaches the population
limitations: can’t use for multiple, not good for larger problems, not good for >1 metric
mixed integer progamming for facility location: obj function, constraints, variables
why used MIP instead of cross-median?
obj function: min distance, min travel time
constraints: sum of zi = L (number of locations)
* can also specify a range of L
* can also specify locations that can’t be opened simultaneously
variables: binary zi = 1 if you open a facility in location 1; 0 otherwise
possible approaches to multi-objective optimization
- weights -> can be hard to work with when you’re trying to decide what weights to use
- penalties → same as weights; if an objective has to be min, use a negative weight
- turn the objectives into constraints
* can plot objectives against each other, ex. revenue vs fuel
* looks like an efficiency frontier
* choose where you want to be
nearshoring, reshoring, protectionism
nearshoring — offshoring to a nearby country
**reshoring **— back to the homeland
protectionism — protect national economic interests; less open movement of goods/labour
* can be oriented around geopolitical relations
how to create an ethical climate, in general and from an ops perspective
- start w company values
- empower employees at all levels to make decisions
- don’t punish mistakes
ops perspective:
1. metrics and motivation
2. understand the system/demand so that employee expectations are realistic
3. innovate, be responsible leaders
issues with patient scheduling
urgency classes and criteria
surge capacity
scheduling within a day, ex. super urgent case
breakdowns/cancellation by provider
system/software downtime
patient no shows by urgency class
possible patient scheduling rules
recommended metrics?
- first available slot
- protecting level: reserve capacity for most urgent cases or for each class
- Patrick et al.’s rule: fill tomorrow (lost anyways), then book as late as possible without exceeding target; if exceed target, use overtime or surge capacity
- least busy day prior to target
recommended metric: Proportion of patients of a specific priority class who receive the
service within a specific clinically desirable time
LP, IP, MIP + how to make dec variables binary
linear programming — all variables continuous
integer programming — all variables are integer variables and/or binary variables
* binary variables: constraint > bin
mixed integer programming — some combination of continuous and integer/binary variables
what is the obj of a yield management analyst? how does that differ from that of a sales rep?
yma: maximize aircraft utilization (revenue per passenger and load factor) by approving/denying group booking
sales reps: set fare and preserve customer relations with corps doing group bookings
yield management, revenue management
segmenting the flights market?
yield management — ideal operating strategy for companies that face temporary imbalances between capacity and demand, spoilage; enables companies to maximize use of constrained productive capacity with a discriminating eye on product yield
revenue management — selling to the right customer at the right time for the right price, with the right options, right distribution channel
fare class, advance/last-minute, dist channel, time of day, route
3 yield management tools
where is the tradeoff? up to what point do you oversell?
overbooking → passenger reservations > capacity; accounts for no-shows, last-minute cancellations, and missed connections
* increases options for passengers
* generates incremental revenue
* starts off high 6 months before departure and slowly declines
discount allocation → spread fare classes over seat sections in 1 plane; saves seats for higher-valued, last minute business customers
traffic management → allocates discounts, valuing high-paying passengers more, looking at the network of connecting flights
overbooking vs spoilage is the tradeoff; oversell where MC = MR
