Supply curve: servers for your restaurant

Data-Driven Decision Making for Business

Ted Kwartler

Data Dude

Determine an appropriate supply to meet the demand

Approach 1

Inputs

  • 100 calls in 30 minutes
  • 5-minute transaction time

$$

Extreme outputs

  • 100 agents
    • no customers wait
    • 5 minutes of work per employee
    • 25 minutes of idle employees
Data-Driven Decision Making for Business

Determine an appropriate supply to meet the demand

Approach 2

Inputs

  • 100 calls in 30 minutes
  • 5 minute transaction time

$$

Extreme outputs

  • 1 agent
    • 94 customers wait
    • 30 minutes of work per employee, non-stop work
    • Some customers would just hang up! Missed opportunity!
Data-Driven Decision Making for Business

Erlang-C formula to the rescue

  • Four inputs
  • Returns the number of agents/servers/workers needed

$$

Input Definition Example
Rate How much demand is occurring? 100 calls per 30min
Duration How long to complete 1 unit? 5 minutes per call
Target How long can work wait? People can wait up to 20 seconds
Service goal What percentage of work units need to be serviced under target? 80% of calls will be serviced under 20 seconds
Data-Driven Decision Making for Business

Math? Ain't nobody got time for that

$$

Erlang-C

Data-Driven Decision Making for Business

Balancing incoming calls, wait time, and agent occupancy

Inputs

  • 100 calls in 30 minutes
  • 5 minute transaction time
  • People can wait up to 20 seconds
  • 80% of calls need to be answered within 20 seconds

Results

Erlang_results

Data-Driven Decision Making for Business

Hurry up and wait

Data-Driven Decision Making for Business

Preparing Video For Download...