Power Advocate, Inc. Confidential 1 Why 80/20 is really 50/50 The Energy Spend Data Story
Jun 21, 2015
Power Advocate, Inc. Confidential 11
Why 80/20 is really 50/50
The Energy Spend Data Story
Power Advocate, Inc. Confidential 22
At PowerAdvocate, we’ve cleansed more than $1.7 Trillion of energy spend data
Along the way, we’ve seen an almost universal divide between what energy executives are led to believe their data looks like…
….and what their data actually looks like
Power Advocate, Inc. Confidential 33
For supply chain organizations, this divide is costly…
Lost operational efficiencies
Millions of dollars in lost savings
Lack of spendvisibility
Power Advocate, Inc. Confidential 44
In this presentation, we’ll show you actual statistics and real-world examples from energy companies’ data to illustrate…
What Energy Spend Data Looks Like
Why Quality MattersWhy It’s So Bad
Power Advocate, Inc. Confidential 55
Initial State of Spend Data
What Energy SpendData Looks Like
Why Quality MattersWhy It’s So Bad
So, what does the data look like?
Power Advocate, Inc. Confidential 66
There’s the perception:
Visible and Actionable
Not Visible
Non-PO
P-Card
Recurring Services
Stock Materials
Recurring, Non-Stock, and
Capital
Non-Recurring, Non-Stock, and
Capital
“I have an accurate view of 80% of my spend”
What you think is 80/20 is more like 50/50
Then there’s the reality:
Typically,50% of this
is sourceable
…and this consists of scattered,
“dirty” data
Power Advocate, Inc. Confidential 77
43%are not
of transactions
internally classified…
On average,
Power Advocate, Inc. Confidential 88
How would you source these categories?
of classified spend data
a. “services – other” b. “services – general” c. “services – misc”
…and 40 - 60%ends up in one of these
3 categories:
Power Advocate, Inc. Confidential 99
duplicatessuppliers are
(that should be rolled up to parent companies)
1 in 4
Supplier Project ID Description
Rosemount Inc 00292486 EMERSON 3051 TMT P# 03031
Alliance Cooling 00214539 Phase II Upgrade
Clariant Oil Services 00252302 2600 LTR TEG 30/70 POD
CEMEX 00431123
transactions lacksa line description…
2 in 5
Plus… and…
Power Advocate, Inc. Confidential 1010
MILLIONSfeed into more than of annual transactions
3 incompatible systems
ERP APT & E
P-CardLegacyExcel
Not to mention that
Microsoft Excel reaches its data limit at ~ 1 Million rows
For reference:
Power Advocate, Inc. Confidential 1111
The averaged statistics are pretty bad; let’s see how things look at the individual organization level with real-world case studies
In each case, supply chain executives were entirely unaware that the following problems existed within their data:
Lack of Completeness
1Poor
Organization
2Lack of
Granularity
3Inaccuracy
4
Power Advocate, Inc. Confidential 1212
Top 20 Categories
Status Quo Best-in-Class
This customer had ~$1.5 Billion within it’s ERP classified as “(blank)”
(blank)
$1,457,050,525
Customer Classification:
No Classification 101 Categories
Lack of Completeness
1
Power Advocate, Inc. Confidential 1313
GE ENERGY INDUSTRIAL SOLUTIONS $ 6,053,332 GE MOBILE WATER INC $ 1,784,292 IGE ENERGY SERVICES UK LTD $ 357,940 GE MOBILE WATER INC $ 352,936 GE INSPECTION SERVICES INC $ 330,372 DRESSER INC $ 236,306 GE ENERGY EMISSIONS TESTING $ 130,186 BENTLY NEVADA INC $ 125,860 GENERAL ELECTRIC CO $ 65,979 GE DIGITAL ENERGY $ 61,720 BENTLY NEVADA INC $ 44,805 GE ENERGY ALTAIR FILTER TECH $ 28,006 DRESSER INC CONSOLIDATED $ 27,512 GE INFRASTRUCTURE SENSING $ 25,705 BENTLY NEVADA INC $ 20,541 GE OIL AND GAS INC $ 20,496 GENERAL ELECTRIC CO $ 18,129
Status Quo Best-in-Class
41 GE Subsidiaries / Spend Breakdown$9.7M GE Spend - $9.7M
1 Vendor
This customer had 41 GE subsidiaries classified as separate entitiesPoor
Organization
2
Power Advocate, Inc. Confidential 1414
Pipeline Construction Services $456,304,461 Compression Construction Services $174,432,534 Station/Plant Construction Services $171,408,614 Pipeline Maintenance and Repair $72,564,112 Mainline Construction Services $58,585,944 Offshore Construction Services $33,482,086 Engineering $20,500,387 Pipeline Painting and Coating $17,809,263 Site Work $12,034,333 Electrical Construction $10,878,110 Excavation $9,125,016 Meter & Gate Station Construction Services $8,544,243 Station/Plant Maintenance Services $8,257,971 Rotating Equipment Maintenance $6,038,837 Inspection Services $4,818,307 Environmental Services $3,650,723
Status Quo
1 Category 186 Categories
Services > Construction > Other
$1,133,762,876
GL Account Description:
Best-in-Class
This customer classified over $1B as “services > construction > other”Lack of
Granularity
3
Power Advocate, Inc. Confidential 1515
$58.9 M
Initial Engineering Spend
Spend that Wasn’t Actually
Engineering
Spend that was Engineering, but Was Classified Elsewhere
The customer had no visibility into $80.4M that wasn’t effectively sourced
As a result, it failed to uncover massive opportunities for savings and improved efficiencies
-$11.0 M
+$32.6 M
Best-in-ClassStatus Quo
This customer’s engineering spend was badly misclassified and scatteredInaccuracy4
Power Advocate, Inc. Confidential 1616
Think you might have a “status quo” data problem?
…Or read on to discover the root causes of spend data problems
Click Here, and We’ll Show you How to Fix it
Power Advocate, Inc. Confidential 1717
Why It’s So Bad, and Why it
Matters
What Energy Spend Data Looks Like
Why Quality MattersWhy It’s So Bad
Let’s consider why the data’s so bad in the first place…
Power Advocate, Inc. Confidential 1818
People in the field are responsible for operations,
not spend classification
Power Advocate, Inc. Confidential 1919
They have no training, time, or incentive to select purchase codes that mostly closely reflect transactions…
When faced with materials and services masters that contain thousands of options each…
So they choose ‘safe’ options like: “services –
general”
Power Advocate, Inc. Confidential 2020
And with hundreds of your employees frequently making requisitions…
Misclassifications happen thousands of times every day within your organization
Power Advocate, Inc. Confidential 2121
The Result? By now you know…
ERP data that’s highly un-actionable due to a lack of completeness, organization, granularity, and accuracy
Power Advocate, Inc. Confidential 2222
Or, in other words…
ERPSystem
Garbage In Garbage Out
Power Advocate, Inc. Confidential 2323
But why can’t your data be fixed on the back end?
Surely with enough analysts and elbow grease the job could be done…
Power Advocate, Inc. Confidential 2424
Let’s take a look at 1 transaction…
Power Advocate, Inc. Confidential 2525
This is what you bought:
Power Advocate, Inc. Confidential 2626
7IN 26.00# 0.362W P-110 HC ERW R3 LTC-HYUNDAI-REG MIL(175 JT)
And this is what it looks like to your analyst:
Power Advocate, Inc. Confidential 2727
It’s a bit complicated…
7IN 26.00# 0.362W P-110 HC ERW R3 LTC-HYUNDAI-REG MIL(175 JT)
Curious how he would decode and classify this?
Power Advocate, Inc. Confidential 2828
7IN 26.00# 0.362W P-110 HC ERW R3 LTC-HYUNDAI-REG MIL(175 JT)
Manufacturer
Hyundai Automobile
Hyundai Railway
Hyundai Steel
Hyundai Ship
Hyundai Plant
Hyundai Electric
…
He would likely recognize “HYUNDAI” as a Hyundai Corp manufacturing facility, but that still leaves hundreds of plausible classifications
?
Power Advocate, Inc. Confidential 2929
Material Group
Shape Steel
Wire Rod
Tubing
Piping
Casing
Alloy Ingot
Manufacturer
Hyundai Automobile
Hyundai Railway
Hyundai Steel
Hyundai Ship
Hyundai Plant
Hyundai Electric
… …
So, he would need to know that the welding specification “ERW” denotes casing, tubing, or piping; this narrows material groups, but leaves 3 options
?✓
7IN 26.00# 0.362W P-110 HC ERW R3 LTC-HYUNDAI-REG MIL(175 JT)
Power Advocate, Inc. Confidential 3030
Material Group
Shape Steel
Wire Rod
Tubing
Piping
Casing
Alloy Ingot
Manufacturer
Hyundai Automobile
Hyundai Railway
Hyundai Steel
Hyundai Ship
Hyundai Plant
Hyundai Electric
… …
Outside Diameter
12 Inches
7 Inches
5 Inches
Third, he would need to know that “7IN” indicates an outside diameter, and that this measurement rules out tubing; still, 2 material groups are possible
✓
7IN 26.00# 0.362W P-110 HC ERW R3 LTC-HYUNDAI-REG MIL(175 JT)
✓
?
Power Advocate, Inc. Confidential 3131
Finally, he would need to know that “0.362W” references thickness, and that an item from this facility and material group with these dimensions is OCTG Casing
Material Group
Shape Steel
Wire Rod
Tubing
Piping
Casing
Alloy Ingot
Manufacturer
Hyundai Automobile
Hyundai Railway
Hyundai Steel
Hyundai Ship
Hyundai Plant
Hyundai Electric
… …
Outside Diameter
12 Inches
7 Inches
5 Inches
Thickness
0.250 Inches
0.362 Inches
OCTG Casing ✓
7IN 26.00# 0.362W P-110 HC ERW R3 LTC-HYUNDAI-REG MIL(175 JT)
Power Advocate, Inc. Confidential 3232
This single transaction could have been classified in more than 200 different ways…
See how difficult accurate energy spend classification is?
Power Advocate, Inc. Confidential 3333
Now imagine that 1 transaction…
…multiplied by millions (50% of which is unclassified, blank, and/or inaccurate)
It’s a decoding job not fit for a human
7IN 26.00# 0.362W P-110 HC ERW R3 LTC-HYUNDAI-REG MIL(175 JT)
Power Advocate, Inc. Confidential 3434
Out of curiosity…
Let’s see what it would take a human (your analyst) to cleanse this data to a 98%+ level of accuracy?
Power Advocate, Inc. Confidential 3535
x 7-1/16" 5M, w/2 2-1/16" 5M S, Cameron C
Tubinghead, 11" 3M x 7-1/16" 5M, w/2Scrubber, 24" x 4', Horizontal, 300 # MAWPScrubber, 24" x 4', Horizontal, 300 # MAWPve-Gate, 1-13/16", 10M PSI, Cameron FLS, FE
Valve-Gate, 1-13/16", 10M PSI, Cameron FL
An analyst would need to cleanse 312 transactions per hour (or 1 transaction every 12 seconds) for a full year to cleanse just 1 year of data!
1 average year of data = 1.3 million transactions
Assume (conservatively), ~ 50% of transactions – 650,000 lines – require cleansing and classification
With no vacation, there are 2,080 hours in a work year
Which means…
Here’s the math:
Power Advocate, Inc. Confidential 3636
That’s right – he would need to decode one of these:
every 12 seconds, with no breaks, for a full year(and that would just get you one year of cleansed data that is a year out-of-date)
7IN 26.00# 0.362W P-110 HC ERW R3 LTC-HYUNDAI-REG MIL(175 JT)
Power Advocate, Inc. Confidential 3737
Now do you believe your spend data could be better?
…If the answer is “yes,” but you’re not sure why you should care, read on
Click Here to See How Much Better
Power Advocate, Inc. Confidential 3838
What ‘Good’ Looks Like
What Energy Spend Data Looks Like
Why Quality MattersWhy It’s So Bad
But why should you care about quality?
Power Advocate, Inc. Confidential 3939
12% Savings
$…which, on average, amounts to millions of dollars every year
on the spend you could be actively managing, but aren’t
You fail to realize
Source: Aberdeen Group
Power Advocate, Inc. Confidential 4040
Assume you have $1B of sourceable spend
With better visibility, you could manage an incremental $500M by improving sourcing and reducing off-contract spend…
And achieve 12% savings as a result
Which means…
For Example:
You leave $60M potential savings on the table every year
$1B $1B
+$50
0M
Status Quo Best-in-Class
Visi
ble
and
Man
aged
Visi
ble
and
Man
aged
Power Advocate, Inc. Confidential 4141
And, to make matters worse…
You’ve hired a team of expensive analysts…
Power Advocate, Inc. Confidential 4242
Data scientists spend 80% of their time mired in “data janitor work.”
Who end up with no time to analyze!
Power Advocate, Inc. Confidential 4343
There’s another cost that hurts your bottom line
The average, entry-level data analyst makes $60,000/yr.
Let’s assume you have a modest team of 5 analysts, who spend 80% of their time cleansing spend data rather than analyzing it
Which means…
Let’s look at your best-case scenario:
You spend (at least) $240,000 every year for analysis you never receive
Power Advocate, Inc. Confidential 4444
Ready to join the world’s top energy supply chains?
Discover How to Achieve Best-in-Class Data