Analytics on your inventory
Friday, July 13, 2012
By doing an automated analysis of your inventory levels, you can work out ways to increase your chances of having the part you need readily available, whilst reducing the amount of money tied up in inventory, says ASCI.
By applying statistical analytics tools (such as the ones developed by Oniqua) to your inventory, you can achieve a rare combination reduced stock outages and lower cost of inventory ownership at the same time, said Scott Hawkins, President, Chief Operating Officer and co-founder of Advanced Supply Chain International, speaking at the Digital Energy Journal Stavanger conference on Feb 28th, 'optimising supply chains'.
Many people think optimizing inventory only means reducing the amount of inventory, but often the more useful improvements are in making sure you have important parts readily available when you need them she aid.
If you have the right parts available, and so stock outages and maintenance delays waiting for deliveries, 'at the end of the day you save money,' he said.
Another benefit of doing automated analytics on inventory is that you have a method to resolve disputes about whether a certain part should be stocked, or taken out of stock.
'New items keep coming. People will say, 'it would be nice if you carried this'. It is hard to say no.'
'The maintenance community carries a lot of clout and sway, and in the absence of good rigorous analytics and quantifiable answers, it becomes an emotional conversation - 'I need this' - 'no you don't' 'yes I do, you don't know anything about maintenance'.
'That conversation happens 100 times a month.'
But if you have a computer analytics program to tell you the best answer, 'it makes it easier to say no if you don't need to stock the item and easier to say yes if it does,' he said.
If everybody agrees on the rules, such as the stocking philosophy and criticality classifications, and if there is an understanding of the power of statistical tools, then it 'builds a tremendous bridge' between materials and maintenance organisations, he said.
It is important to explain to maintenance staff how the classification system works, so they understand that important items are more likely to be in stock.
A common error is to neglect including the maintenance organisation in the inventory criticality classification process. If the materials organisation just does the classification in a vacuum and fails to educate maintenance staff on the power of the tools, 'the buy-in that's so essential doesn't occur.'
Once those bridges are built, maintenance managers begin noticing a real difference in inventory performance. For example, if the statistical tools drive warehouse service level (availability of needed parts) from 95 per cent to 99 per cent, that is an 80% reduction in stock outages. The maintenance staff will definitely notice.
This means that out of 1000 orders, the numbers out of stock has reduced from 50 to 10.
'And of the remaining ten stock-outs, the items which were out of stock are far more likely to be the less consequential items, because of the way the available usage forecasting tools treat critical items.'
Once items have been purchased for a capital project or even a major turnaround, if they are not there is often a strong push to get the warehouse to bring them into stock.
Here again, the analytics tools can settle the question. with a few clicks of the mouse, tools such as Oniqua Inventory Optimiser can work out if you are better off paying the costs of returning items, including restocking fees, or if you should go ahead and bring the items into stock based on usage history for like items.
Analytics tools can also help foster better working relationships between maintenance staff and materials staff, when maintenance staff find that the parts they need are more likely to be immediately available.
'We're in the early years of a revolution in business analytics,' he said.
'I think over the next 10 years you'll see this type of thing become much more widespread.'
Unless attention is paid to it, inventory tends to grow over time, much like mould, Mr Hawkins says.
'Once you have it, it tends to grow at a certain pace if you don't have a program to keep it in check'
Stocking levels are often planned around a certain level of usage for the item, and if the item starts to be used less, there is no process to correspondingly reduce stocking levels.
'Eliminating items is a lot more difficult to do in most organisations than bringing new ones in,' he said.
It is also common that oil and gas companies don't have a plan for dealing with excess inventory. Oil and gas people are commonly very good at making plans and putting them into action, but not so good at circling back around to check on status and then optimise=, Mr Hawkins says.
'Around supply chain, what we've observed, is that [planning and executing] get most of the time and energy and there isn't much left over for optimising, checking and re-optimising.'
Yet the oil and gas industry has made a lot of effort to optimise other areas of its business, including reservoir engineering and exploration techniques.
A lot of the barriers to doing optimising work are with master data quality, because you can only optimise a supply chain if the master data is good, he said.
Master data, including catalogues and equipment lists, varies a great deal in quality.
Equipment bills of materials, the list of maintenance parts a certain piece of equipment requires, also need to be accurate, otherwise the wrong parts get ordered.
'Without at least a modicum of master data quality it is very difficult to optimise,' he said. 'It is very difficult to know what you've bought repetitively, to tease out the opportunities, when you have a catalogue rife with duplicates and poor descriptions.'
There are many agencies around the world which specialise in helping sort out master data, he said. 'There are firms which have sprung up which can help you put proper taxonomy and nomenclature in place, do an initial data scrub, and even help maintain the data over time.'
The data quality does not need to be perfect to generate enormous value, he said. The most important pieces of information are the purchasing history of items and past order lead times.
Another barrier to adopting analytics is the technology tools. Many companies spent large amounts of money in the 1990s and early 2000s installing ERP systems and putting all of their data into them, with a view to making this system the only system they used.
But many ERP systems don't have very good data analysis tools, or other specialised analytics tools, and companies are reluctant to use software outside their main ERP system. That is a barrier, says Hawkins, because no single system can do everything. Specialized functions such as advanced analytics rely on ERP system data and functionality, but can greatly enhance the value and efficiency of ERP systems.
Supply chains will not optimise by themselves. And ERP systems won't do it, either. It 'takes some new energy, some external third party stimulation, and some leading edge tools.' he said.
ASCI, or Advanced Supply Chain International, is mainly a business process outsourcing company.
It has a sister company focusing on asset management services, including catalogues, equipment bills of materials, and technical aspects of enabling and operating supply chains, with a focus on optimisation.
The company employs around 250 people, with a further 75 involved in joint venture projects in the Caribbean and Australia.
It helps companies which do high volume transactions, companies making thousands or hundreds of thousands of purchase and receipt transactions per quarter.
If all of your stock has a defined criticality level, that is useful for making decisions about stocking, he said.
The criticality can be defined as the cost of not having the item when you need it, he said.
'A' criticality could be for an item which, if not available when required, would lead to the platform being taken out of action.
'E' criticality is for items which would be inconvenient to be without.
'Hands in the field wouldn't like it, but there would be little cost in not having it,' he said.
You need expertise from maintenance personnel to classify items in terms of criticality.
The calculation of whether a part should be kept locally in stock is a function of the criticality level, the amount you use the part (stock issues per year), and the costs of looking after the stock for each item, including storage, handling, capital tied up in inventory, and other costs.
You can plug this data into algorithms which can tell you the optimum stocking levels for your warehouse, and even flag vendor-held stock opportunities.
There are statistical algorithms which have been around for decades to analyse different types of inventory, he said, with different profiles for items.
By processing the data you can get an answer which 'everybody can agree on,' he said.
Some companies still try to use spreadsheet models to do tasks like this, but these can get terribly time consuming when you have tens of thousands of items. The bottom line is that old fashioned spread sheet models lack the robustness and productivity benefits of leading edge analytics tools.
Behind other industries
Other industrial sectors, such as manufacturing and retail, have been doing optimisation of inventory for many years, and are a long way ahead of the MRO ('maintenance, repair and overhaul') sectors, he said.
'Oil and gas is one of the last industries to embrace this.
'I have my own theories about why,' he said. 'Industries with tighter operating margins have more incentive [to optimise]'.
Managing inventory is not a one-off process - you have to continually re-examine inventory and keep up with staff training.
Once the system is in place, it is important to run annual training in how to use the system, or otherwise, as people move into new roles, 'this process, like any other process, can wane,' he said.
Mr Hawkins recommends that you re-run the inventory analytics algorithms monthly so you can identify, every month, items which are in stock but don't need to be; which items need better vendor management; and recommendations for items which would be better stored in a different location (onshore vs offshore), based on their criticality and usage patterns.
You can identify items which have seasonal use patterns and set stock levels to adjust accordingly.
You can also look at where the largest changes are taking place in your spare parts consumption and then examine them in more detail.
The system will also match the supply to current demand, not the demand 2-3 years ago. 'Demand patterns change a lot,' he said.
One company ASCI is involved with, described as a 'tier 1 multinational oil company,' worked with ASCI to help optimise inventory.
The company had varying levels of master data quality in different parts of the world.
The oil company employs a highly regarded specialist in inventory analytics, who was 'able to get out front and pilot this process and bring his peers along,' he said.
The company started with pilot implementations of the analytics in the Arctic and Indonesia.
'That allowed the organisation to learn some lessons to define some templates around setting up the cost model, and to validate the return on investment (of the effort,' he said. 'Is the view from the top worth the climb?'
The effort for Alaska took six months, and the 'results were eye popping,' he said. The results for Indonesia were 'equally impressive'.
So the company decided to make the effort globally, running on a global analytics platform.
The company's internal studies for the first year showed a 20:1 ratio of benefits to costs for the project, including the costs of the software, set up, software maintenance and implementation.
It will continue to get the same results every year. 'Having your inventory properly optimised on a monthly and quarterly basis is a gift that keeps on giving,' he said.
'The value was not limited to stock reduction; it was also about increasing stock, particularly high criticality and medium criticality items,' he said.
'Having the correct items you need when you need them, and having process to eliminate items you don't need - generates quantifiable value,' he said.
There were also intangible benefits from the improved trust levels between materials and maintenance departments.