Lab Study 2 : KPI and Dashboarding
piggy_cute
MGT153-Session4KPIMetrics1.pptxThe Rady School of Management
Metrics, KPIs and Dashboards
1
Today
Metrics and KPIs
Dashboards
Applichem Case Debrief and Lab
2
What is a metric?
A metric is anything you measure as part of the business
Distills raw data (lots of numbers) into a single number
Typically a summary statistic (average, median, standard deviation)
Ex:
Point-of-Sales Transaction Data -> Total Sales per week ($ / week)
Click-stream data -> Average # of website visits per week
Order invoices -> % of customers that place a second order within 6 months
Be specific! Units matter when defining a metric.
3
What is a KPI?
Key Performance Indicator
KPIs are a type of metric
KPIs embody a strategic objective for the firm
Distinction between KPIs and metrics is subtle
If it doesn’t tie to a strategic objective, but is still useful, it’s a metric
Ex.
Apple: Cost-per-unit-produced ( $ / unit )
Netflix: 6-month Customer attrition rate (%)
You’ve been taught dozens of metrics/KPIs in your time at Rady. Examples?
4
Why track KPIs and metrics?
“What gets measured, gets done”
Sets up incentives
Communicates goals to other departments
Communicates strategy externally
Can guide decision-making
Ensures firm “on-track” with its goals
Measures progress
Alerts management to potential problems early
5
Just because something can be measured, doesn’t mean it should be.
What are qualities of a good KPI or metric?
(Some debate…)
6
Qualities of Good Metrics: Simplicity
Immediately understand / guess how it was calculated
Straightforward interpretation for the business
Bad Examples.
Avg. Credit Score of Loan Applicants: 645
# of defective products returned per month: 35
7
Qualities of Good Metrics: Comparative
MGT4
Financial Accounting
MGT136
Advanced Accounting
Is this number good or bad?
Benchmark this number against
Previous performance
Others in industry
An established target
Requires metric is based on credible, reproducible data
If similar data aren’t available next month, how can I compare?
Ex: Annual Release-Ease Production by Plant (Millions of lbs)
| Gary | Frankfurt | Mexico | Sunchem |
| 14 | 38 | 17 | 4 |
8
Qualities of Good Metrics: Influence Behavior
MGT133
Advanced Cost Accounting
MGT134
Federal Taxation - Individuals
Most importantly, changes in a metric should shape behavior
Are there things you can do in your business to affect this number?
What will you do as a business if this number goes up unexpectedly? Goes down unexpectedly?
Examples of “not so influential” metrics:
Number of sick days per employee
9
Summary
Good metrics/ KPIs are
Simple
Comparative
Behavior shaping
It knowing this number won’t help you achieve your businesses goals, don’t bother tracking it.
10
Aggregation and time frame
MGT45 Principles of Accounting
MGT181 Enterprise Finance
MGT166 Ethics & Corporate Responsibility
MGT112 Global Business Strategy
MGT16 Personal Ethics at Work
KPIs and metrics that are useful to some members of the business are useless to others
11
Aggregation and time frames (II)
Entrepreneurship
Generally
Higher management -> more aggregated, long-term metrics
In the trenches -> more granular, short-term metrics
Ex.: Car-Dealership
CEO of GM -> Avg. Quarterly Sales
Director of Regional Sales -> Avg. Monthly Sales by Car Model by Dealership
Head of Dealership -> Weekly Sales by Car Model by Salesperson
12
Your turn:
Entrepreneurship
Organizational Behavior
Marketing
Finance
Operations
Are the following good KPIs to track?
Most popular hours of the day for riders
Most popular hours of the day for drivers
Average wait-time (in minutes) a rider waits for a car
Average surge pricing paid
Number of times individual drivers sign-in to the app
Average rating of drivers
% of times a user switches to Lyft (a competitor) because of a long-wait
You run the LA operations for UberX.
Minor is customizable
See sequence map for list of which electives fall into which categories
13
Data requirements
Always consider what data is necessary to compute your metrics
Do you already collect it?
How frequently?
Periodically? In real-time? Only by special request?
When in the transaction is it collected?
Point of sale? Point of delivery?
By whom is it collected?
Automatically? Back-office? Internationally? In multiple distinct places?
Often, collecting the exact right data is too costly
Can you proxy your metric by something simpler for which you do have the data?
Requires care
14
Exercise:
Return to the previous Uber exercise
What are the data requirements for each metric? Think about
Do you have all the data you need?
Where are they collected currently? When?
If you don’t have the necessary data, can you proxy this metric with something else?
15
Dashboards
16
Businesses Are Viewed from Four Key Perspectives
The Balanced Scorecard—Measures that Drive Performance:
https://hbr.org/1992/01/the-balanced-scorecard-measures-that-drive-performance-2
Case: Electronic Circuits Inc.
The Balanced Scorecard—Measures that Drive Performance:
https://hbr.org/1992/01/the-balanced-scorecard-measures-that-drive-performance-2
Most quantitative analysis can be performed with charts that use only four kinds of objects
Selecting the right visualizations…
Change over time
Comparisons
Ranking
Part-to-whole
Distribution
Correlation
Overview
Geographical information and Maps
What is a dashboard?
Dashboards present a small set of KPIs/metrics through a simple, easy-to-understand interface
Might be numbers and words
Might be a visualization
Usually a combination
Dashboards provide a rapid, overview of the organization at a glance
Some people only use the term “dashboard” for real-time information, visualizations, or displays that fit on one page.
They use “report” otherwise.
We will use them interchangeably.
21
What are dashboards used for?
Overview the business rapidly
Educate departments about important objectives
Provide alerts when problems occur
22
What makes a good dashboard?
Intuitive, easy-to-understand
Aesthetically pleasing
Contains only the right metrics
Consider your audience. What decisions do they need to make?
Minimize extra details
Presents data around a theme to guide decisions and create a story
Puts metrics in context. Is this number high, low, out of the ordinary?
23
The quintessential Dashboard
What are the pros/cons of this design?
24
Is this a good dashboard?
Source: http://www.kaushik.net/avinash/digital-dashboards-strategic-tactical-best-practices-tips-examples/
25
What about this?
Why?
Source: http://www.kaushik.net/avinash/digital-dashboards-strategic-tactical-best-practices-tips-examples/
26
A better example
27
Psychological elements of dashboard design
Visual aesthetics are more than just “looking” pretty
Actually affect the way we understand the data
Information Overload
Mental fatigue
Selective Attention Tests
28
Applichem Case Debrief
29
Applichem’s business model
Revenues driven by single flagship product : Release-Ease
International supply-chain and factories
International customer base
| Sales | Plants | Actual production | Exports by region | Imports by region | |
| North America | 29 | Gary | 14.0 | 0 | 15 |
| Western Europe (incl. Middle East and Africa) | 20 | Frankfurt | 38.0 | 18.0 | 0 |
| Latin America | 12.3 | Mexico | 17.2 | 4.9 | 0 |
| Pacific and Rest of World | 11.9 | Sunchem | 4.0 | 0 | 7.9 |
30
Competitive landscape
How do the US and European markets compare?
What sorts of impacts do these differences have on the business?
| Market Characteristics | Europe | U.S. |
| Competition | Fierce | Low |
| Quality standards | Closely monitored | Normal |
| Typical Shelf-life | <1 yr | Up to 3 years |
| Size | 50-kilo | from 1/2-kilo on up |
31
Comparing the plants
Gary Plant
8 formulations, 80 package sizes
19 other product families
1,000 non-union employees
Capacity: 26 million lbs / yr
Frankfurt Plant
Technologically advanced plant
12 other product families
600 employees
Capacity: 47 million lbs / yr
Mexico Plant
Exclusively sells 50-kilo bags
6 other product families
Less educated workforce
Capacity: 27 million lbs / yr
Sunchem Plant
Many 1/2-kilo and 1-kilo packages
High energy costs
Capacity: 5 million lbs / year
32
Why are these differences between plants significant?
33
KPIs for Applichem
Management spent considerable expense and time
Sharing best practices to improve efficiency
Reassessing capacity allocations across plants
Reallocating production to match demand, labor costs and foreign exchange risk
After all that, they want to make sure what’s fixed doesn’t break again
As part of the operations team, your task is to design a set of KPIs to track to alert management of significant changes in
Productivity
Labor Costs
Efficiency
34
Understanding data requirements
Open the .csv file “Production.csv”
This file contains the last year of data that is currently tracked at Applichem
Can you calculate your KPI from this data?
(Don’t actually calculate, just assess if you can.)
If you cannot, what additional data would need to be collected? Can you proxy your original KPI?
35
