Category Archives: Decision Optimization

Re-introducing Keelvar, An Optimization Backed Sourcing Platform

Last summer, we introduced you to Keelvar, a provider of an optimization-backed Sourcing Platform with a strange name that produced uncommon results. A client of Keelvar’s with 30 year’s experience in ocean freight sourcing remarked that it was like their first generation sourcing suite on steroids, so Keelvar is definitely a platform that should be on your radar.

A spin-out from a research laboratory that built a next generation solver, Keelvar was formed as a SaaS (Software as a Service) company in 2012. Unlike the majority of sourcing optimization software providers, they realized from day one that the solution wasn’t a better optimization module, it was a better sourcing platform where ease of use was the most important driver of adoption. Their Founder and CEO, Alan Holland, believes that “every sourcing event with a possible split-award should be optimized but in a completely transparent manner so that it’s easy for both buyers and suppliers” to understand the award recommendation and the reason therefore.

The starting premise was that the average user wasn’t advanced enough to use the first generation strategic sourcing decision optimization (SSDO) solutions where optimization was a standalone module and you needed to be an optimization platform expert to set up the models and run the sourcing event. It had to be intuitive and easy. Moreover, these average buyers have much simpler categories. As a result, the models were simpler than what some of the first generation expert providers were throwing at them, and, moreover, the majority of these models for standard, “simpler“, categories were fairly standard and could even be mapped to a default template by a senior buyer that the junior buyer could start with, and sometimes use unaltered.

Unlike many platforms, the optimization-backed sourcing platform walks the buyer through a simple process for creating and executing a sourcing event that even a junior buyer will feel comfortable with, detailed in our introductory Keelvar post. This process is as easy to use as a powerful, weighted, RFX tool or e-Auction, which is something that cannot be said for the majority of optimization solutions on the market that still require a solid understanding of the underlying mathematics and operations research. It’s sourcing for the every-man.

And it’s not just for the mid-size company anymore. If you will recall, when SI first reviewed the solution, SI identified it as the perfect solution for the lower-end of the mid-market. However, over the last six months, the usability and power has been steadily improved, and now SI will also identify it as a perfect solution for the entire mid-market and a candidate solution for the Fortune 500 / Global 3000. Keelvar, which has a US presence in addition to its UK presence, has acquired a number of Global 3000 clients for whom the solution is perfect for. For the majority of categories, and spend, in these organizations, there is no need for overkill optimization and the solution is a perfect fit. That’s not to say that it can’t handle the more complex categories either, with a number of global logistics events for some of the world’s largest shippers already under its belt, just that it is perfect for those enterprises where the majority of events are not that complex and can be templated for the average buyer.

Plus, the Kevlar platform is under continual development and improvement. Not many SSDO platforms can say that either. When optimization is one module in a set, providers can’t focus constant development on it. When optimization powers a single platform, it’s a different story.

Any company that chooses Keelvar gets a solution that allows all spend to be subject to the appropriate level of optimization, which is what allows Keelvar to get fantastic results for some organizations and why they have progressed so fast. If you have a traditional, heavy, solution that can only be used by a small set of dedicated resources on a few large categories, the organization will struggle to get 10-20% of spend under optimization-backed management which means, at the end of the day, the savings potential of the solution (expected to save 10%) is a mere 1-2% on the bottom line. But if you have a solution that is lighter, streamlined, and useable by everyone in your organization, you can get 80% of spend under optimization-backed management. And while the simpler categories won’t hide the same opportunity, even if you squeezed out a mere 6.25% average savings, that’s still 5% to the bottom line and a much better investment. At the end of the day, it’s not the line item count that the product can support, but the amount of spend that flows through it. And the bravado claims by first generation solutions that their solution can support tens of thousands of line items have done more harm than good.

Keelvar is an agile company and is developing new product features and global coverage quickly. New developments in the last six months include …

  • RFI with bulk editing capabilities.
  • Bid data sanitizer to detect, correct and prevent erroneous bidding (via automatic identification of outliers).
  • Enhancements to multi-round RFQ capabilities.
  • Large scale e-Auction support for hundreds of lots and bid upload capabilities.
  • New support portal to cater for the growing global customer base.
  • Partnerships with benchmark data providers in Global Logistics.
  • A partnership with an e-Sourcing Suite provider.
  • New offices in South America and Australia (which give it a presence on four continents).
  • Paid trial offering for Global 5000 companies (and, as per a recent SI post, paid pilots, under the consulting budget, could be just what you need to prove the value of an optimization-backed sourcing platform and get investment $$’s to upgrade your platform, and your results).

As one of the few companies aggressively developing an optimization-backed sourcing platform, and one of the few companies in the Supply Management space focusing on adoption first, Keelvar is one of a handful of global companies that might just lead Procurement out of the dark ages (in which Procurement seems to perpetually exist as a resident of the Island of Misfit Toys) and into the enlightenment.

Mass Adoption of Optimization via Modern Sourcing Platforms

In a previous post, we addressed three common misconceptions about sourcing. In this post we expand upon those corrections we provided to give you six pillars of a properly designed optimization-based sourcing platform.

The three pillars of a properly designed optimization-based sourcing platform that we addressed in our last post were:

  • Useable by everyone, including the most junior of buyers
  • Affordable as any other sourcing platform
  • Efficient, decreasing event set-up time by at least a factor of 2 to 3

In addition, an optimization-based sourcing platform is also:

Powerful

Optimization is powerful. A modern optimization engine can solve sourcing problems to 99.9%+ optimality in a matter of minutes, even if they require tens of thousands of variables and hundreds of thousands of equations to describe. The platform is effectively evaluating hundreds of thousands, or millions, of different award splits in a matter of minutes.

Valuable, more so than any other sourcing platform

Simply put, optimization gets amazing results. Even if the category has been negotiated repeatedly over the last ten years, and it looks like the savings opportunities are razor thin, with the ability to analyze more suppliers, more bids, more transportation options, more value-add options, more constraints, and more supplier-specified opportunities, optimization can often identify an additional savings of 10% or more. In fact, the year-over-year average savings from optimization alone on the categories it has been applied to has been clocked at 12%, more than any other sourcing platform.

Insightful

With optimization, you can create different scenarios, with different suppliers, constraints, and goals and see how the optimal awards differ as the problem definition changes. This inspires a sourcing analyst to look at the problem in different ways.

However, the first three statements in particular are only true if the platform used by default is an optimization-backed sourcing platform . Classically, optimization solutions have been implemented as stand-alone platforms. These were powerful, and when used by the right senior resources who set up the right sourcing events, these platforms generated amazing results, but they were very difficult and time-intensive to use compared to an e-RFX or e-Auction platform. The model had to be set up. RFX data had to be imported. The data had to be validated and cleaned. The model was then run, altered, and re-run until an acceptable baseline was found. Then multiple what-if scenarios were run until a final award scenario was identified. Then the award scenario had to be exported to the sourcing platform so the suppliers could be notified and the contract(s) drafted. All of this was very time consuming. As a result, the platform was not useable across the sourcing organization, was not very affordable as it had to be supplemented by other sourcing platforms, and this process was definitely not efficient.

For mass adoption of optimization, it needs to be supported by an RFX and/or an e-Auction platform for data collection, by analysis and reporting for result presentation and exploration, and needs to be integrated with contract management and the supplier portal for negotiations and communication. In other words, optimization is the engine that powers the modern sourcing platform, it is not a stand-alone solution.

That’s why a modern optimization-based Sourcing platform, and not a standalone optimization module, is the silver lining that Procurement has been waiting for. What does this platform look like? Stay tuned!

Optimization: What’s changed since 2009?

If you’ve done your research you have likely figured out that it requires a PhD to use optimization. That it is so expensive it should only be used by experts for high value categories. And that the time required to set up optimization for a sourcing event prevents you from using it for more than a handful of events.

All of the above statements were true in 2009. But these statements are completely false today, and have been false for a while now.

Let’s take them one by one.

Optimization requires a PhD.

Modern tools do not require a PhD, a Masters or even a Bachelor’s Degree. A properly designed optimization solution should be usable by junior buyers as it should hide the complexity and simply provide one-click evaluation. With built-in rules, workflows, wizards, templates, and other modern usability features, optimization can be as simple to use as an e-Auction on an average category for a junior buyer.

Optimization is very expensive.

While it used to be the case that optimization solutions were expensive, sometimes costing six (or even seven) figures for a single event, this was pre 2010. Today, mid-sized organizations can receive 40 events with unlimited users for that or unlimited events for a small number of users. Take your pick.

Optimization makes a sourcing event very time intensive.

Modern tools have intuitive web-based workflows to make event creation a task conducted in minutes. While it might take a long time to set up a complex event with dozens of suppliers, hundreds of line items, thousands of lanes, and hundreds of constraints, not all events are that complicated. Many categories are only bid to a few suppliers, need to be shipped using the two or three existing contract options, and don’t have a lot of constraints. This allows the model to be set up, the data imported, and a baseline solve completed in a matter of hours, not days. Moreover, if the model is set up as a template, it can be copied and reused over and over again, and setup is a matter of minutes. Done right, optimization decreases the amount of time it takes to set up a sourcing event by a factor of two or even three or four.

In other words, optimization has become mainstream and should be considered a default strategy for all sourcing events. And what does a modern solution look like? That’s the subject of our next post.

Optimization Backed Sourcing Platform … Or Bust Part V


This is the fifth and final part of a five part series that revises and ties together key ideas outlined last year on Sourcing Innovation across multiple posts. Regular readers will be familiar with much of the content, but the integrated perspective should help to cement the ideas in regular readers and new readers alike.

This post is largely based on It is NOT Direct or Indirect — It is Strategic and Complexity!.

In this series we’ve been discussing the need for a modern optimization-backed sourcing platform by first explaining how it’s not a suite, it’s just sourcing, then explaining how it’s not sourcing, but strategic sourcing, and, finally, that it’s not just strategic sourcing but optimization-backed strategic sourcing in order to master the complexities of the modern global supply chain.

We’ve also spent a significant amount of time going into detail as to why first generation suites, which were typically nothing more than a loosely coupled set of pseudo-related modules organization around a theme, did not make a modern platform, and certainly not one that could be called a modern optimization-backed sourcing platform for strategic sourcing of complex categories across modern global supply chains.

But, as per our last post, there is still one more argument against the need for a modern optimization-backed sourcing platform that we need to address. Specifically, the argument that only custom-manufactured direct categories are complex enough to need a complex solution and that the average CPG or indirect category won’t benefit from an overkill solution.

This argument is flawed. And rather than try to attack it, we’re going to get right to the point. The right way to source a category has absolutely nothing to do with whether it is a direct category for your organization or an indirect category for your business. Nor does it have anything to do with whether or not it is a category regularly sourced by your GPO or whether or not the GPO has it under contract.

Remember, as per our last post, even the categories that were traditionally seen as the simplest indirect categories are sometimes actually among the most complex “direct” categories that the organization possesses! In SI’s paper on Complex Sourcing: Are You Ready, we elucidate how even a category such as paper is among the most complex categories that the organization can source.

Secondly, what is indirect for your organization is direct for another organization, and a supplier in particular. Calling it indirect only masks the fact that, at some point in the supply chain it is a complex direct category and if your supplier, or GPO, is not approaching it correctly, a significant amount of money is being left on the table.

While there are some that would very much like to forget that before the introduction of e-Negotiation (e-RFx and e-Auctions), a number of “indirect” categories used to cost organizations millions — such as tires in automotive, lights in aviation and printer ink in back offices everywhere — this is not the right thing to do. We have to remember that these organizations never understood how much these “secondary” categories were really costing them and that, sometimes, 100% profit margins were the norm, because they often did not have the ability to go out to market like we do today.

Thirdly, while a product organization might see services as indirect as such a category would be labelled as non-core, and, similarly, while a service (or financial) organization might see a product category as indirect as it too would be labelled non-core, if such service, or product, is essential for the organization to deliver the product, or services, the organization profits on to the end consumer, how can such a service, or product, really be non-core?

For example, if successfully selling that next generation cellphone requires augmenting the supplier’s design team with a new design team that can enhance usability above the competitor’s product without sacrificing a low-price point or quality, that is a critical service and should not be treated as a secondary outsourced indirect category. Similarly, if delivery of your big data analytics services requires a specific high-end laptop configuration that can not be easily met by all providers, and a sub-par configuration would result in delays or service degradations, this is not a category that can be thrown over the wall to a GPO either.

In other words, direct or indirect has no correlation to the complexity of a category or its strategic importance to the business and, thus, should not be used to determine the appropriate sourcing strategy or sourcing platform. The right way to initially classify a category is to use a basic measure that that captures its strategic importance and its complexity and any category with a measure that exceeds a certain threshold must be strategically sourced. The rest can be sourced using simple spot-buys or other traditional methods provided that they are not too complex, or too strategic in someone’s view, for these traditional methods.

And, as such, there is no argument not to use an optimization-backed sourcing platform on every single category that the organization needs to source.

Optimization Backed Sourcing Platform … Or Bust Part IV


This is the fourth part of a five part series that revises and ties together key ideas outlined last year on Sourcing Innovation across multiple posts. Regular readers will be familiar with much of the content, but the integrated perspective should help to cement the ideas in regular readers and new readers alike.

This post is largely based on Why Your First Generation Sourcing Platform Is Not Ready For Modern Sourcing.

As per our earlier posts, first generation sourcing platforms, circa 2005, were a miracle cure for the average Sourcing organization that was drowning in data and demands to save, save, save without enough time or resources to tackle even a fraction of the categories that needed to be under management.

For example:

  • First generation Spend Analysis systems helped the Sourcing team identify the largest spend categories and the largest organizational suppliers, which were prime candidates for the first strategic sourcing evens put through the new sourcing platform.
  • First generation RFX systems helped the Sourcing team capture more data from more suppliers than ever before and not only better qualify potential suppliers but collect more detailed bid breakdowns for analysis.
  • First generation e-Auction systems helped the Sourcing team put non-strategic high-dollar categories with very little complexity out to bid for quick savings success.
  • And, most importantly, first generation decision optimization systems allowed the sourcing team to build realistic cost models, capture constraints, and devise realistic award scenarios that identified real savings.

In other words, many organizations that acquired these suites and applied them successfully saw year-after-year returns of 10%+ on the spend brought under management. And a few are even seeing some savings today, but just like the second auction saw little savings and the third auction saw a price increase, the year-over-year return is dropping. Why? Because while these first generation platforms were infinitely more powerful than anything that had come before, they weren’t designed to capture the full extent of complexity in an average category — complexity that has been considerably increased since the early days of sourcing due to increased outsourcing, increased globalization, increased regulation, and a constantly evolving global marketplace. Back in the day, no one would have thought that even a simple paper tender (that’s right, paper) would encapsulate all nine dimensions of sourcing complexity as defined in “what defines complex sourcing and why does it matter” on Spend Matters.

Not only is there just no way to do a strategic analysis and justify a strategic decision without a basic level of true mathematical optimization capability that can take all costs and constraints into account, but there is also no way in a standard suite to:

  • Collect the full breadth of supplier responses
    as the limited form-based data collection in first generation solutions don’t allow for expressive responses, bids, or varying from a rigid, inefficient, yes/no fixed cost template
  • Create dynamic drill-down graphical reports
    as the built-in static reports with limited, if any, 2-d graphing options don’t allow for modern graphical analytics
  • Conduct multi-user sourcing events
    as they were set up for non-collaborative single-user sourcing events and not modern, complex, categories that require entire teams
  • Create accurate cost models
    as the limited cost models, with little or no formula support, are too primitive for many of today’s complex categories
  • Support dynamic workflows
    that adapt to the categories and the organizational needs, as everything has to fit the rigid workflow or the suite cannot be used

A modern sourcing platform is needed, and by now it should be quite clear that it should be optimization-backed, but there is still one common argument that is used by many vendors to try and skirt the issue, and while it’s not valid, it can sure be convincing when spoken by a charming and charismatic salesmen — and this is the topic of the fifth and final post in this series.