Category Archives: Manufacturing. Metrics

Product Price Pandemonium? Covalyze and Clarify!

In a nutshell, Covalyze (by Valunoo) is an advanced analytics platform designed to help an engineer and/or a direct buyer understand the price quotes they are receiving from a supplier by helping them understand what the average price is, what the (target) price of a similar product should be, and what factors are contributing most to the product, regardless of what type of product you are considering.

While primarily used for manufactured products (with no restrictions on categories), it is not restricted to (select categories of) (electro) mechanical or electronic products like some platforms. Nor is it restricted to products where you have unit prices, as it also encodes the necessary algorithms to analyze bundles (which are common in software, for example) and if you have enough quotes, understand the underlying pricing of each module.

The platform, founded and run by two economists and built by German trained mathematicians and data scientists, encodes advanced curve fitting, regression analysis, cluster analysis, support vector machines and k-nearest neighbors and other advanced statistical and machine learning algorithms to analyze all of the available commercial data against the relevant technical details of the product and its production process.

The platform can be used standalone if you have, or can export, all of the data into decently formatted csv flat files, or it can be used in conjunction with their partner platforms that will digitize technical drawings and/or specifications, collect available market data on raw material production and pricing, cleanse and classify all organizational data in the organization’s systems to ensure the data used for analysis is accurate and complete enough to be used, and to integrate external commercial data feeds for market intelligence. (They and/or their consulting partners can resell these platforms so you only need to deal with Valunoo or your preferred consulting and implementation partner.) Covalyze has partnerships with ChAI, KNIME, Konfuzio, and Werk24 that it can use to power it’s new Data Connector (for Super Analysts) which it can deploy to help you build your perfect data pipeline. It also has partnerships with catuuga, EFESO, and P3 (management) consulting firms that can provide you with the services you need to help you integrate your systems, cleanse your data, and get all of the product data into the platform you need to do your cost analysis (and support your negotiations).

Feature Based Price Breakdown

The Covalyze platform revolves around the calculation of the zero margin price line using a feature-based price breakdown that will allow the engineer and the buyer to understand

  1. whether each price/quote from a supplier has a positive or negative contribution line (and you want to minimize a supplier’s positive contribution margin as that means you are paying more than average and not optimizing your pricing) and
  2. what features of the product are contributing the most to the price/quote

Once all of the data for all instances of a product (from multiple suppliers and or quotes) with their relevant features and costs are loaded into the platform, the platform runs its suite of analytics, selects the appropriate algorithm, and computes the feature-based price breakdown for each instance of the product and then the zero-margin line. (Note that, since all of the techniques are essentially statistically based, you want at least 30 for decent accuracy, and if you have 100, or 1000 instances, load those too … some clients have up to 2000 instances of a product, although you will find negligible increases in accuracy beyond 100 in most cases).

At this point, the buyer can

  • plot all of the products/quotes against the price line and see how many are above/below the line
  • dive into the feature-based pricing overview for the product and see the average contribution of each “feature” to the product (which would be gross area, coating, edge type, stamping, material, thickness, and supplier variance for a sheet metal part, with most of the cost dependent on the gross area and the rest split among the other features)
  • dive into a specific supplier product/quote and see the cost breakdown by feature

Once they understand the price line and feature-based price breakdown, they can dive into the

  • savings potential
  • target price calculation
  • feature price calculation

Savings Potential

Once the price-line has been computed, if the platform is also fed, or has access to, the data on which products are currently being bought, in what volume, and/or the forecast data, it can be used to compute the cost savings if:

  • all suppliers making a positive margin contribution reduce their pricing to the the zero-margin line or
  • all products that can be replaced with a lower cost product from another supplier are replaced with the lowest cost alternative

Target Price Calculation

Not only can you see the average price for a product across multiple instances, and see the average contribution of each feature, but you can see the price impact if you change (or remove) a feature. For example, if different products used different material options (steel, aluminum, titanium), you can see the impact on the target based on changing the material type.

Moreover, as long as you have the data, you can create your own customized part configuration (by selecting your own options for all of the features) and the platform will tell you what the target price should be based on all of the data at its disposal.

Feature Price Calculation

Not only can you see how much a feature type contributes to a target price in the target price calculation, but you can see how the price varies across the different options, and if those options can be mapped to a scale (like area, length, weight, etc.), then it can extrapolate the expected (target) price of that feature based on available data.

Similar Parts

If you need to find a replacement part (because a supplier suddenly becomes unavailable or the particular part you were using becomes too expensive), or you don’t have enough data on a feature in a part to fully understand the cost or the options, you can use the similar part explorer that will show all similar parts in a category with one or more features in common.

Model Builder

The platform allows the user to construct feature-based cost models for any product of interest within the platform, which will then simplify the loading of data, or allow the user to target cost a new product currently being designed (as long as there are enough products in the database with cost information on each feature of interest, where the features do not have to be in the same, or even similar, products).

Discussion

Covalyze doesn’t do “should cost” because that depends on market and private data you can never have access to (as labour rates are not just country and region dependent, it’s company dependent — if the company is using state of the art technology on its production line, it may need better trained engineers which will cost more; if a company is close to a hydroelectric plant, it might be able to cut a private deal lower than average energy rates in the country that still has a lot of oil burning plants; if it has it’s own water supply or water processing facilities, it’s costs will be significantly lower; etc.), but it can do very informed target cost as you can determine what data to include in the baseline calculations. You can limit to products/parts that only use a specific material type, are produced in a certain region, only use a certain production technology, etc. and create a highly defensible target cost model that you can use as a baseline in your negotiations with suppliers, which can help you get suppliers to disclose their real cost — because you can say “this cost is based on the products we buy (or have quotes for) that use only these materials, are produced on this type of (production line) technology, and come from factories in your region, so if your costs are wildly off, you need to tell us, and tell us why“. Not to mention, you probably already have tools that allow you to build should cost models, but don’t have any access to the data needed to populate those models. So you can start with Covalyze target costs and force suppliers to justify increases.

One key point is that Covalyze is NOT just for Procurement Intelligence — it is great for Sales Intelligence as well. Nothing stops a sales organization from going to competitor and distributor websites with price listings and detailed product specifications, loading that data with their product data, computing the price lines, and seeing how their product will place against their buyer’s zero-margin line. They can also see which costs dominate in the cost models, which features are demanding a premium price, how they can defend their pricing if it is above the line, or how they can claim that a competitor’s product is overpriced. Insight is the key to success!

In the hands of an astute and acclaimed analyst, you’ll quickly be able to get at the true cost factors of any product, including one you haven’t sourced yet (which could be a manufactured product, a SaaS suite, or even produce [as it can compute cost differentials between farms, greenhouses, and hydroponics against lifespan and waste costs if you so desire]) the true performance of a supplier against a market, and the true target cost baseline you should be seeking, with exceptions only made with concrete evidence from the supplier that the desired quality, detail, etc. can not be obtained at the market price (and that, thus, all other products at market price will be inferior for your needs).

The implementation timeline is dependent only on three things:

  • how long it takes you to arrange system access to all of the systems that data needs to be pulled in from (API keys, access to data lakes where there is no API and just data dumps, etc.)
  • how many parts you want to model
  • how many part instances in total you want to pull in

This is because:

  • until you can access all of the data (which will be spread across the PLM, the ERP, the Supply Chain S&OP, the Sourcing, the Procurement, and the Finance systems, you can’t do anything but define the part models; if you have a slow/uncooperative IT, or archaic systems, it could be weeks before Valunoo or their consulting partner can start the implementation
  • models depend on the complexity of the product (they are just a list of features with associated types and, in the case of an attribute set, defining the attribute sets); they can be created pretty quick from scratch for many parts by an experienced engineer, and both Valunoo and their partners have starting libraries that they can use to customize a model for your business (but these are not included out-of-the-box as the factors you include are dependent on what data you have, how significant a cost factor has to be to be meaningful to your organization [because if it only contributes about 1% to the cost of a $100 product and you only buy 10,000 units a year, and the savings potential is max 10%, you’re not going to do a deep dive analysis for a max cost saving of $1000 — you’d spend more in computing power building the model than you’d end up saving])
  • it takes a lot of time to initially populate each part instance because you have to collect data from multiple systems; do multiple levels of conversion, cleansing, and validation; and then run multiple types of analysis to identify the best cost model (and run it); on average, it’s 8 hours to load 50 part instances and do all of the initial analysis and computations; so if you have 50 types of parts with 100 parts each, that’s about a month with standard provisioning (and while you could provision a larger virtual server and parallelize multiple streams, this would cost more but your analysts probably couldn’t keep up with the manual analysis anyway; but if you were a large enterprise with 500 parts and 50 to 2000 instances a part, you would likely provision multiple instances (for completely different product types) and then large instances for big categories

This being said, most clients will be up and running in a month, with all parts and data integrated within a quarter.

Covalyze is a great analytics solution for design engineers and direct buyers who really want to understand the pricing dynamics of the current and potential supply base, which quotes are truly good or bad (against a baseline), the features (and requirements) that are really driving the price (which is not always material), and what changes, if made, should substantially lower the price. It’s the type of solution every Direct Sourcing team needs to have in their repertoire.

Will Trump’s America First Policies Put America Last?

Trump wants to bring production back to America, and that’s a noble effort and, for many companies, a smarter thing to do than they realize as escalating logistics costs and global uncertainty make near-shoring and, even better, home-shoring much less risky (and, in the long run, often more cost effective) than off-shoring, especially when there’s no good reason to off-shore.

But Trump’s recent almost across-the-board tariffs are going to cost some American manufacturers anywhere between millions of dollars to hundreds of millions of dollars as, simply put, due to a lack of availability of certain resources, Americans have to import. The net effect of so many lower-cost global options over the years is that American companies went off-shore for just about everything they figured they could get cheaper, and as a result not only has there been little to no growth in raw-material extraction and production at home, but some industries have actually lost capacity. And that capacity can’t be turned on and ramped up over night.

As a result, Americans need to import aluminum, steel, and other metals, at least for the short term. And while most of that importation should come from near-source locations (like Canada and Mexico, especially if the US wants to maintain NAFTA, which, for the most part, is better for it than Canada and Mexico [combined]) to decrease risk and increase border security (after all, it has two borders — Canada and Mexico; working with Canada and Mexico on security issues makes the entire North American continent safer), Americans have such high demand in some categories even Canada and Mexico can’t meet it all.

For now, American manufacturers have no choice to but import their raw materials from other (non-exempted) countries. It’s unfortunate, but it’s the reality. And if any of these companies have access to good global strategic sourcing optimization and supply chain planning tools, they’re going to start modelling and realize that it’s cheaper in the mid-term, and maybe even the short term, to manufacturer whatever is intended for the global market outside the US. Rev that factory back up in Mexico and serve the world from there. Only manufacture at home what is needed at home.

And what happens if companies shift their operations to other jurisdictions? America loses jobs, tax revenue, and it’s share of the global GDP. That’s, hopefully, not what Trump, or anyone inside North America, wants.

And while there should be tariffs on goods imported from jurisdictions a country can’t compete with and, in particular, a country that allows its corporations to pay it’s employees $2 a day for a job an American would have to be paid at least $58 a day for (as there’s no way America could compete with imports otherwise), those tariffs should be designed not to hurt the manufacturers who depend on raw materials they can’t get at home, or at least be used to fund local raw material extractors / producers to give those companies at home a local option. For instance, all tariffs collected should go into a fund to help local raw material extractors and producers expand or increase production, and until that happens, companies that need to rely on imports in the interim should at least get tax credits until such a time as they have a local option. Or they are just going to find ways to take as much of their business as they can elsewhere.

And that won’t make America great again, or even competitive. While I actually agree with the premise that, especially when it comes to manufacturing and agriculture and staple industries, America needs to be great again, unfortunately, just slapping import tariffs without a broader plan to achieve that goal is not only not going to help, but it’s going to hurt.

Procurement Trend #09. New KPIs

Only six sinsational anti-trends to go. That means we’ll reach the end of this series in two weeks. We should be all happy, happy, joy, joy! But how can one be joyful when one realizes that this means we had to slog through a two-four of anti-trends to get here, and that some of us probably had to drink a two-four in the process to keep the dark thoughts at bay! And, to be honest, the doctor is really worried that there is no skin left on the futurists’ drum (which has taken more of a beating than any drum set Neal Pert has ever owned) and that, giving their predilection for ancient trends, these futurist historians may try to skin LOLCat to give their drum new life.

So why do these funky* futurists keep trying to push new KPIs as a future trend? Is it because of their fondness for three-letter acronyms that stems from their party-hardy frat-days (filled with a little too much beer pong)? the doctor has his suspicions, but it’s probably because they finally figured out that:

  • what gets measured gets managed is still trueand KPIs require measurements
  • new processes and new technologies mandate new measuresso KPIs need to be updated whenever a new process or technology is brought in, which should be a regular occurrence in a best-in-class Supply Management organization that makes an effort to keep up with the times
  • new measures provide new opportunities for improvementjust like new Intel cores have provided new opportunities for faster computation ever since it was all about the pentiums

Measure to Manage

If the only reason that you are measuring spend, year-over-year changes, and captured savings is to report those metrics at the monthly meeting, then you are doing it all wrong. If you are not using base measurements as a foundation to identify inefficiencies and opportunities for improvement and repeated measurements as the foundation for evaluating progress, then you shouldn’t be measuring in the first place. You’re better off spending your time in old-school hard-nosed negotiations because, at some point, you might actually whip that sales rep so hard that he forgets which way is up and goes under the floor just to escape the verbal onslaught. (Of course, you will create disdain in the supplier who will do the bare minimum to fulfill the contract terms and, if the rep buckled too much, you might even bankrupt the supplier who hopes to make it up on future deals but never does, but hey, at least you got those impossible savings, right?)

Measure to Master

It’s not enough to measure just to track the status and success of current initiatives, you should be measuring with a goal of achieving mastery. If the benchmark for the average throughput in your industry is 100 invoices/day/clerk, then you should be striving to get your exception-based invoice automation process to 100/day/clerk error-free invoices and nothing less should do. If you don’t get there in the projected amount of time, you should be introducing new measures that break down, or influence, the process flow such as resolution time per exception invoice, average buyer response time per clerk contact, average number of line items on a problem invoice, etc. until you figure out where the slow-down is and what you should do about it. (Automatically reduce exceptions by kicking invoices back to suppliers with explanations of errors and do not allow resubmission until corrected, mandated response within 48 hours or a black mark in the buyers’ performance review, break down POs to insure more manageable invoices, etc.)

Measure to Excel

This means not just measuring process, throughput, and savings but finance measures favoured by the C-Suite, even if they do not help Supply Management directly improve performance. At the end of the day, if Finance is happy, the C-Suite is happy, and Supply Management is much more likely to get the financial resources it needs to implement new systems and processes that will ultimately improve the metrics even more.

* and not the good kind of funk

Open Up Your Supply Chain With E2Open

Today is the official launch of E2Open‘s new Collaboration Center, E2Open Version 8.0. The focus of this release are their new supply dashboards with real-time KPIs, predictive analytics and exception notifications designed to allow an organization to manage its global trading network across multiple supply tiers.

E2Open was founded in 2000 with the vision to provide supply chain managers visibility into their entire supply chain network — beyond just the first tier of suppliers because problems often start with your suppliers’ suppliers and your suppliers’ suppliers’ suppliers. Getting visibility into a late shipment or raw material shortfall as soon as it happens gives an organization time to find an alternate supply or alternate go-to-market strategy, as opposed to finding out the day after your supplier was supposed to ship. Since then, E2Open has gone through multiple versions of its platform and its E2open Business Network (8 to be precise) and now offers solutions in Collaborative Supply Planning, Demand Management, Logistics Visibility, Order Management, Inventory Management, and B2B Managed Services with a customer list that includes Blackberry, Dell, FoxConn, Hitachi, Motorola, and Seagate to name a few.

However, today we are only going to focus on its new collaborative platform and its supply management dashboards to be precise. Why would I do such a thing, especially since I repeatedly claim that Dashboards are Dangerous and Dysfunctional in full agreement with Robert D. Austin? Because the reason they are dysfunctional is that they lull you into a false sense of security when you see a lot of green. As I said in SI’s now classic post:

a dashboard can not tell you how well you’re doing … the best it can do is capture the data it’s been programmed to capture, roll-up the metrics it’s been programmed to roll up, and do the built in calculations of efficiency based on those roll-ups.

As a result, even if it tells you that 90% of spend is “on contract”, that doesn’t mean it is. It won’t tell you that 10% of spend has been misclassified under the wrong code and is being reported as on-contract when it’s really, really not. The truth is that:

a dashboard can only provide an upper bound on how well you’re doing, and this is useless. Reporting that my efficiency is at most 98% when it is in fact 92% is useless and unactionable.

However, if the goal is reversed from trying to tell you how well you are doing, and giving an inaccurate upper bound, to how poor you are doing, and give an accurate, minimal lower bound, it becomes useful. And if you can then define metrics such as inspected orders, reviewed invoices, verified shipments, etc. and report on the uninspected orders, unreviewed invoices, and unverified shipments (etc.), then you not only know everything that’s wrong but how many dark corners could be holding problems waiting to materialize but where to look when the problems you know about have been solved.

And that’s why E2Open’s new dashboard, developed in HTML5 and available through your browser, is useful. Not only does it provide deep, near real-time insight into your global supply network, with data aggregated across the multiple tiers of your supply network as fast as the platform can get access to it (which is real-time if the suppliers are using a modern supply management system with real-time query / export capability or once a day if the supplier is still on an old ERP/MRP that does a daily export in CSV to a secured FTP directory), but the drill-down dashboard can be configured to display whatever KPIs and metrics you want, however you want.

You can choose the standard indicators that show that 98% of your orders are expected to ship on time, based upon tier-1 and tier-2 suppliers shipping their components and raw materials on time, or you can invert it and show that 2% of your orders are late. Every metric can be reversed and you can filter what is displayed. So, if you want, you can set it up to show ALL RED and just show you

  • all the problems the system has identified that need an investigation and/or resolution and
  • how many records, products, shipments, etc. have not been manually reviewed, tested, verified as this will tell yo exactly where problems could be lurking and, if the count is high, where more oversight might be required to prevent new problems.

It’s not the standard configuration, but it is supported — and the ability to razor sharp focus into issues two levels down into your supply chain within 24 hours of your supplier’s supplier reporting a delay is fantastic. And, unlike most “dashboard” products, they support the creation of multiple public and private “dashboard” pages, at different levels of visibility and granularity, to allow each user to track all KPIs, metrics, and issues relevant to them. It’s not trying to be a one-size fits all solution because E2Open recognizes that, in supply chain, one size does not fit all.

Furthermore, 90% visibility at each tier is possible very quickly as they have done over 400 ERP / MRP / Supply Chain system integrations to date and can on-board suppliers on all of the major platforms very quickly. And they even have the ability to do trending and predictive analytics to identify where problems might occur — which is useful when you know that somewhere in a certain data blackhole there is likely an issue but are unsure where to start.

E2Open’s new release is worth checking out. The platform strives to give you a single version of the truth across your supply network and does a good job at doing it. And the inventory management / collaborative forecasting drill down capability is just as detailed as some of the best inventory solutions on the marketplace.

Key Takeaways from the UL Product MindSet Study, Part II

A couple of posts ago, we discussed some Interesting Facts and Figures from the UL Product MindSet, a recently released study that quantitatively surveyed 1,195 manufacturers and 1,235 consumers across a range of export and import markets in high-tech, building materials, food, and household chemicals. Then, in our last post, we reviewed four key takeaways from the UL Product MindSet Study. Today we are going to discuss our fifth, and final, takeaway from the study.

MANUFACTURERS NEED TO GET A GRIP ON REALITY!

They need to take off those rose-coloured glasses, put them on the floor, and stomp them to bits. And then they need to take the bits and grind them into dust. The findings illustrate that manufacturers are so far out of touch with reality that it’s downright scary.

First of all, let’s review the standard Gaussian curve. In a standard curve, only 31.8% of the population is one standard deviation from the norm. If we accept that only one standard deviation from the norm is enough to be “ahead of the curve”, then, at most 15.9% of the population can be ahead of the curve (and, similarly, 15.9% of the manufacturers will be behind the curve). However, the report found that an extreme majority of manufacturers believed they were ahead of the curve in safety, reliability, sustainability, and innovation. In short, this means that:

  • 81.1% of manufacturers are out-to-lunch when it comes to product safety
  • 81.1% of manufacturers are day-dreaming when it comes to product reliability
  • 78.1% of manufacturers are high-on-fumes when it comes to sustainability
  • 73.1% of manufacturers don’t-have-a-clue when it comes to innovation

The reality for the majority of manufacturers (68.2%) is that, they are, at best, on the curve. But since the reality is that, if they don’t continue to progress as their supply chains evolve around them, it won’t be long before them are behind the curve, they should just assume they are behind the curve, because 15.9% of them are and 68.2% of them aren’t far from being among that 15.9% without continued improvement efforts. So when they are done grinding those rose-colored, haze-inducing, glasses into dust, they need to get to work!

Furthermore, I see no evidence that the majority of manufacturers understand sustainability. I know it’s hard with all the greenwashing out there, but if one just ignores the hype and uses a little common sense, one can define sustainability as that which sustains operations and the environment at the same time. With this definition, it is easy to see that if an organization is not reducing its environmental footprint and at least maintaining, if not increasing, profitability at the same time, it is not sustainable. So 69% of manufacturers are wrong when they say that environmental products aren’t profitable — because, defined (and designed) right, they are.

And those manufacturers who do understand some of the basics of sustainability obviously don’t understand it’s importance. First of all, it’s not just about sustaining the environment, its about sustaining operations for generations to come. If the resources available are depleted before they can be replenished, there’ll be no materials to make new products. No products, no profit. No profit, no business. It really is that simple. As a result, sustainability should be as important as safety and reliability, not only one-fifth as important. Secondly, with even the majority of consumers in developing countries (such as China where four-fifths of the population would buy a truly green product over a non-green product if proof of claims could be provided), an organization is leaving what is potentially the biggest gold-vein available to it untapped. And finally, if manufacturers as a whole don’t change their understanding and their views, then the lot of them are are being hypocritical! (It is impossible to be ahead of the curve in sustainability, as 94% of manufacturers ridiculously claim to be, while not placing the same importance on sustainability as is placed on safety and reliability.)

Yes this is harsh, but face it, manufacturers are not going to move forward if they continue to believe the all-rainbows-and-roses picture that some other misguided (or is that money-grubbing?) analysts are painting for them. But there is a bright side. Whereas a typical organization would probably pay five, or six, figures for that rainbows-and-roses report, this post is 100% free.
(So, to any manufacturer reading this, stop calling me a downer and get to work! If you do, maybe you’ll be one of the 15.9% that is truly ahead of the curve and reap the rewards that come from earning that status.)