Category Archives: Best Practices

Time for Spring Cleaning. Start With Your Evergreen Contracts.

The spring tradition is to clean house, and that means your house of business as well. Chances are there a number of areas of your operation that need to be cleaned up, but the place to start is your evergreen contracts. Most organizations have significant overspend in these contracts because prices have dropped, demand has increased, and/or new options have entered the market — but since the organization decided to set, and forget, these contracts, it has not been able to take advantage of new options, negotiate against the increased demand, or realize the reduced prices.

So how do you start?

First, make sure all of your contracts are in electronic form and in a central electronic filing cabinet.

How do you do this?

Acquire a good OCR solution and feed all your paper contracts through it and create a set of contract e-documents.

Then, acquire a good network drive scanner to find all of your e-contracts. Some might be part of the scanned set (as they were printed out and filed), some might be duplicates (as different users might put them on different drives), and some might be draft versions.

Finally, to get a complete (as you can) distinct set of contracts, run them through a semantic process that can identify similar documents that will group all documents that are highly similar into a set and identify the (likely) final version based on dates (and differences between similar documents).

Then, figure out which contracts are, or could become, evergreen …

How do you do this?

Acquire and apply a semantic analytic solution that can sift through the contract clauses, identify the term, and whether or not the contract is, or could become, evergreen.

… and order them by upcoming (auto) renewal date.

This is just a simple sort, which can be done by exporting the contract titles and (auto) renewal dates to a spreadsheet which is easily sorted.

Then do a spend analysis (and projection) on each category defined by the contracts, in (auto) renewal order, and when the savings percentage is significant (near double digits) or the savings amount is significant (many 3X to 5X times what a category re-sourcing would cost), provided there is enough time to re-source, you queue up the sourcing event. If there is not enough savings potential, or time, you add it to the end of the queue to be reanalyzed sufficiently in advance of the next auto-renewal date.

Eventually you’ll work your way through all the evergreen contracts, and replace them (with non-evergreen contracts) in order of priority, defined by savings potential.

And that’s how you start your evergreen contract spring cleaning.

Will the Trade Wars Be Good for Advanced Sourcing?

Trump is imposing tariffs. China is retaliating. And this is just the beginning. As a result, supply risk and the need for spend forecasting is finally becoming real. But is it becoming real enough for organizations to take action? It’s hard to say. But one thing we do know is that the only way organizations can progress forward is to better understand not only the risks, but the costs.

What are the risks? Many. What are the costs? Significant. And how can you know of either? In the first case, you need to monitor the news, the sentiment of the responses in regards to the news, crowd-source some predictions, and run some advanced analytics on all this data to determine the probability of something happening — and sticking.

And in the second case, you build should cost models with current data, and projected data, to determine the impact of a tariff on the total cost of ownership of the product. This means that a simple RFX or Auction platform is just not enough – an organization needs a platform with deep should cost modelling and the ability to create what-if should-cost models based on projected and anticipated changes.

But even that’s not enough. If the projected increases are significant, then the organization will, at the very least, need to reallocate global supply chains to insure that products, which are currently sourced from multiple suppliers and/or locations, are being exported from and imported into the most cost effective locales the organization has access to. And if this is not enough to keep costs under control, then the organization may need to even source from additional suppliers (in different locations) or re-source the entire category (to the extent possible).

But it’s hard to figure all of this out without an optimization backed sourcing platform. Hopefully this is the kicker that is needed to get these powerful analytical platforms into the hands of more Sourcing and Procurement organizations, as these platforms are desperately needed and reduce spend on analyzed categories by an average of 10%+ year-over-year, making their ROI immense.

But, alas, only time will tell. But if bankruptcy could be on the line (when a tariff wipes out the entire profit margin), maybe this time these platforms will finally take hold.

Where are You on Your Master Data Journey?

You want to get cost under control. Maybe even save. You need to ensure compliance. You need to satisfy the auditors. You want to know the risks you face. And the risks you could face. All laudable goals, but all goals that are unobtainable without … you guessed it … data.

More specifically, clean, rich, up-to-date, relatively complete data … which, likely, resides in multiple systems, duplicated across each. This makes data centralization, which is necessary for any of these initiatives, complicated, and often difficult. It’s not just the last update record date, especially since some systems do the last update at the record level, and not the data element level.

Plus, how do you know which parts of which records can be combined? Especially when they conflict or don’t line up. Without an appropriate master data management strategy, and a system that can handle master data management across multiple, loosely related, supply management and enterprise, it can be downright impossible for any initiative that spans more than a few dozen providers or categories. And even that is an effort.

But MDM is not easy to define, and even less easy to implement. First of all, which systems do you use for master data when there is an argument for multiple systems that store a record, such as a supplier, to be a master data system. Secondly, when you do identify the master system, how do you manage, and approve, updates … and how do you insure they get synched to the right systems at the right time? Third, how do you integrate all the data into a single, even if only virtual, record so that you can run a spend report. A compliance report. A risk report. An audit report?

The point is that it’s not just as easy as selecting a system, proclaiming it your MDM, and believing the implementor that your MDM problems will be solved in a few months. Some companies, that aren’t heavily focussed on, and involved with, the initiative take years to integrate systems and arrive at a nearly clean set of master data.

So before you march forward on your next, data intensive initiative, maybe you should step back, ask yourself where you are on your data management journey, and give an honest answer.

RPA: Are We There Yet?

Nope. Not even close. And a recent Hackett study proves it.

Earlier this month, The Hackett Group released a point of view on Robotic Process Automation: A Reality Check and a Route Forward where they noted that while early initiatives have produced some tangible successes, many organizations have yet to scale their use of RPA to a level that is making a major impact on performance, likely because RPA has come with a greater-than-expected learning curve.

Right now, mainstream adoption of RPA is 3% in Finance, 3% in HR, 7% in Procurement, and 10% in GBS – Global Business Services. Experimentation (referred to as limited adoption) is higher, 6% in HR, 18% in Finance, 18% in Procurement, and 29% in GBS, but not that high, especially considering the high learning curve for the average organization will end up with a number of these not continuing the experiment.

Due to the large amount of interest, Hackett is predicting that, within 3 years, RPA will be mainstream in 11% of HR Organizations, a 4X increase, 30% of Procurement, a 4X increase, 38% of Finance, a 12X increase, and 52% in GBS, a 5X increase, as well as increases in experimentation. Experimentation will definitely increase due to the hotness of the topic, but mainstream adoption will require success, and as Hackett deftly notes, successful deployment requirements have certain key prerequisites too:

  • digital inputs
  • structured data
  • clear logical rules can be applied

And when the conditions are right, organizations:

  • realize operational cost benefits
  • have less errors and more consistent rule application
  • benefit from increased productivity
  • are able to refocus talent on higher-value work
  • strengthen auditability for key tasks
  • have enhanced task execution data to analyze and improve processes

But this is not enough for success. Hackett prescribes three criteria for success, which they define as:

  • selecting the right RPA opportunities
  • planning the journey
  • building an RPA team or COE

and you’ll have to check out Robotic Process Automation: A Reality Check and a Route Forward for more details, but is this enough?

Maybe, maybe not. It depends on how good of an RPA team is built, and how good they are at identifying appropriate use cases for RPA, and how good they are at the successful implementation. Success breeds success, but failure eliminates the option of continued use of RPA, at least until a management changeover.

Don’t Get Sucked in By Impressive Words!

It’s conference season, and that means marketing overload for many vendors. And there’s a few words the doctor is hearing a bit too much and he’s NOT impressed! So what are these words?

Digitization

Digital. Digital Procurement. Digitized. Digitized Procurement. Digitization. Ugh. They’ve been using variations of the same word for almost 20 years — and despite claims to the contrary, the meaning hasn’t really changed. You’re analog, or you’re digital. There’s no degrees to digital.

Look at the dictionary definition for crying out loud! Of, relating to, or using data in the form of numerical digits. What’s new, or even enticing, about this? ABSOLUTELY NOTHING!

Internet of Things

The internet has ALWAYS been an internet of things. Computers are not people. They are computers. The only difference today is that we are sticking computers in more things to collect and transmit sensor data automatically rather than reading it, and entering it into the computer. It’s not the big whoop most companies are making it out to be as most companies haven’t developed much that uses that near real-time in a truly useful way.

Cognitive

It’s not artificially intelligent. It’s cognitive. And the bull crap has reached a whole new level. Let’s look at the definition.

Of or relating to the mental process of perception, memory, judgment, and reasoning.

Yes computers can perceive through sensors, store data in memory, use algorithms to assign, or judge, and use very advanced automated algorithms to reason, but we’re overlooking one key word here. Mental. Computers don’t have a mind, and they are not intelligent. The implication here is that which is cognitive is intelligent, and they are not intelligent.

We haven’t even reached true AI yet in any field and we are supposed to believe that a little Sourcing or Procurement vendor has reached the next, cognitive level of AI development? While a best in class vendor may have a few algorithms that are almost cognitive for a few, select, situations, considering the billions going into AI research and the limited progress most specialist vendors are making, you know we’re not ready to be throwing this term around.

And, an honorable mention (because, while not common in our space yet, it’s coming):

Postmodern

the doctor‘s been seeing this word a lot on social media in marketing and commentary, and, unfortunately, it seems like it’s starting to creep into our space. For those of us that actually went to a real University and have a sound (classical) education, we know that Postmodernism is a rather broad intellectual movement across the arts and fields with applied arts (like architecture and archaeology) based on a philosophy that takes us from the literary-influenced philosophy of modernism to a post-modern way of thinking that developed in the middle of the last century and reached wide acceptance in the 1980s, when it was a Land of Confusion.

This was the time of the MRPs (and not the ERPs). Do we really want to be associating our new and innovative solutions with that era?

So please, please, please don’t get sucked in by the the impressive words. Instead look for impressive, time-saving, value-adding functions (and forget the feature lists). (But that’s another rant.)