This weekend, I'd like to welcome Eric Hiller back to Spend Matters. Eric, who is founder and Chief Product Officer at aPriori, will be contributing a series of guest weekend posts on the subject of driving Spend Management upstream. And this is the sixth installment. Please join in welcoming Eric to Spend Matters!
Yesterday we introduced the dimensions of a Cost Model, including methodology. We also discussed a framework (accuracy and precision) to help us relate different Cost Modeling methodologies. If you are coming into this post cold, you might want to read it first. Today, we'll cover the first two methodologies (Comparative Quoting and Expert Opinion), and next week we will finish with Empirical Curve Fits and Mechanistic Models.
Methodology 1 -- Comparative Quoting
The most common method to predictive costing used today (especially in the purchasing function) is comparative quoting – the equivalent of eBay. It's a very old method that has been used since the first caveman who asked two different caveman salesmen for a quote on a stone wheel. It works like this. Purchasing sends the product out for three quotes and two days (online) or two weeks (traditional RFQs) later, purchasing receives three estimates. The basic premise behind this method is that in an efficient market, the market itself will lead you to the true economic price of the part. Therefore, after you take away a customary mark-up percentage, you will have true economic cost to make the part or product.
There are a lot of problems with this approach. First, it is rare that the sourcing agent is bidding in an efficient market. There are only so many manufacturers of automotive axles in the world, especially with enough capacity to serve Toyota or Ford, so the buyer is dealing with more of an oligopoly of suppliers. There is also a significant chance of cartel behavior to develop among suppliers that serve the same customer. In this situation, the suppliers know enough about each others capabilities and capacities and their customers that they help support each others pricing.
Why three quotes? No one knows, but there seems to be a strong tendency that "Three shall be the number of the quoting and the number of the quotes shall be three." However, three quotes are hardly enough to get a real economic price or cost, especially when 'quote variation' or spread is rarely less than 20%. Lots of external factors skew the quotes, including a supplier trying to buy business, a supplier that doesn’t want the business, or if a supplier feels he has no chance to win so his estimate is very rough. Quoting has problems with accuracy, precision, and speed. However, it is very good at moving from target to target; you can quote anything. It is also so time consuming that it creates an inherent barrier to accuracy: We have been told by customers and their suppliers that approximately 90% of all requests for estimates never actually go to bid. This means that the early requests of estimates are not treated with the same attention as final bid requests. Because suppliers know this, the quality of these early estimates tend to be shoddy and of little value to the early decision makers. In addition, the need for early requests cause a cost estimating burden that is ultimately amortized over all the bids that are actually won and, therefore, borne by the customer.
Therefore, besides the fact that comparative quotes are time consuming, from an archery point of view, they are neither accurate (unless you have many of them), nor precise (very unrepeatable). Why not precise? Well, try this experiment: ask for quotes on a part today from ten suppliers. Put the quotes and drawings in your desk for a year. Ask the same suppliers for a quote again. Just to be fair, you can even factor out the variation in material cost per mass changes. How close do you think the year one quotes will be to year two?
Methodology 2 -- Expert Opinion
The first predictive costing method that most companies use is Bob (a.k.a. expert opinion). Bob has been around for 30 years in the company (or at least the industry). He might sit in the cost management group, but he could just as easily be in sourcing or manufacturing. People take designs, plans, and sourcing proposals to Bob for his opinion and cost estimate. He can give one quickly that might not be too accurate or spend a few days and give you a better answer. The human brain is better than any model or computer at being able to skillfully combine experience, disparate cost drivers, and new information and make sense of it. Therefore, Bob can be reasonably accurate on most parts, but his real skill is his ability to re-adjust his costing methodology on the fly to sight on a new target (e.g. a new supplier, new design, etc.).
However, Bob has a lot of trouble with precision. If you ask Bob for an estimate on the same part today and a year from now, you will likely get a different estimate. A manufacturing area manager at one of our customers once explained, "I have three excellent estimators, Bob, Ken, and Betty. I can give them the same part but get three very different answers. And worse than that, I can send the same part to be estimated on the same machine model that resides in two different departments on my floor and get two different estimates of time and cost to make the part. This is killing us. How can we make progress on improving our costing methodology and reducing cost, when we are not even internally consistent?" The problem the manager was experiencing was a lack of precision (consistency).
Bob has a few other problems. Bob does not scale. He has limited capacity. When he retires (as is happening en masse across American manufacturing), it is hard to transfer that learning to the rest of the organization. Holtz-Eakin found that 70% of baby boomers will reach retirement age by 2007. Most of Bob’s hard learned “tribal” costing knowledge will retire with him. In fact, according to Gartner Group, workers spend 60% of their time searching for the tribal knowledge they need to do their work. So Bob is good at what he does, but there is too much work for Bob to do today, and there will be a lot less of Bob around tomorrow.
Both Comparative Quoting and Expert Opinion are very manual processes and have drawbacks and inefficiencies. Their repeatability and consistency is limited. We need some type of machine or 'transfer function' that consistently and repeatably generates product costs. We'll look at two brands of such machines next week.