Never Forget: 80% is Always Enough!

Twenty years ago we wrote a post on why sometimes 80% is enough (when employing decision optimization) because, especially when the alternative is (a) 0% (improvement), 80% can be a monumental improvement.

If you aren’t doing anything, if you can get an 80% improvement with an existing tried-and-true solution, that’s better than nothing. And if you are at 80%, and a tried-and-true solution can get you another 80% of the way to optimal, that would get you to 96%!

While our original post said that “sometimes 80% is enough, sometimes 80% is not even close”, we were wrong because, now that we are in the age of AI, if it will let you avoid an unpredictable BS (Gen) AI solution, 80% is more than enough!

Here’s the thing.

Data on 80% of your spend will allow you to find a lot of opportunities for cost reduction, process improvement, risk reduction, etc.

Data on 80% of your supply options, including carrier options, expected tariffs and duties, defect and waste rates, etc. will allow you to find better allocations than you have now.

Data on 80% of your transactions — where, how, why — personal credit card/expense, P-Card/V-Card, bulk bank transfers from an ok-to-pay in the e-Invoicing system — is enough to allow you to optimize payments and cash-flow.

Moreover, unlike modern (Gen) AI which needs 100% of data in pristine condition to even achieve moderate accuracy, the classical, tried-and-true, solutions that have been serving us well for the past two decades work exceptionally well with just 80% of data available to them.

80% of data allows you to do a much better job vetting suppliers to invite and customizing an RFI than a dumb Gen-AI search solution will do.

80% of data allows you to better judge the right sourcing method — (multi-round) RFP/RFQ, e-Auction, contract (extension) renegotiation, marketplace, GPO/Consortium, etc.

80% of data gives you a great supplier performance profile and allows you to identify if the supplier needs a development program and how best to help them.

In other words, if you can get 80%, in the majority of situations, as long as you avoid (Gen) AI), that’s usually enough to get started (and embark upon a solution that is much better than nothing).

Forecasting Flux

Twenty years ago, we were sure to tackle forecasting and all the potential failures that could arise from the selection of the wrong method or the misapplication thereof in one of our first forecasting posts where we contrasted the two primary techniques in use at the time: judgemental forecasts and statistical forecasts.

In summary, we noted that:

Forecast Type Strengths Weaknesses
Judgemental * Environmentally Responsive
* “Inside” Information
* Responsive to “One-time” Events
* Cognitive Limitations
* Biases
Statistical * Objective
* Consistent
* No Data Limits
* Considers All Variables
* No Real Time Reaction
* Dependent on Model Formulation
* Not Friendly to “Soft” Information
* Technical Understanding Required

The methodology that was typically used in practice was that a statistical model was run and then an expert adjusted it to their liking. Now, this usually worked okay if the organization had a real expert and access to up-to-date soft data, but otherwise, it was a shot-in-the-dark and poor forecasters would saddle the organization with stock-outs (and the ever-present 8% stockout rate) or excess inventory that would have to be fire-sold in the best case, and simply disposed of in the worst case. (Which is very costly from multiple angles, especially if you have to bury 700,000 cartridges in a landfill, which is where Atari’s E.T. found home in 1983.)

Things have changed in the last two decades. Now the average forecaster has two more options, which we really need to discuss: AI and Blended.

Forecast Type Strengths Weaknesses
AI * flexible model selection
* flexible solution methodology
* flexible interpretation of results
* random model selection
* performance limitations
* hallucinations
Blended * dependable models
* multiple scenarios
* balanced solutions that minimize downsides
* computation time
* data requirements
* scenario weighting

The AI option in particular is really concerning. Given that Gen-AI is, at best, artificial idiocy, you have no idea if the model it selects is appropriate, and even if the model it selects is reasonable, AI has never outperformed the classical statistical model equivalent, and then when you ask it for final interpretation, it can sometimes get the interpretation completely wrong. Generally speaking, modern (Gen) AI worsens forecasts.

In addition, the fact that most S&OP platforms still aren’t using balanced blended models is equally concerning. Considering that taking multiple models into account to produce a blended forecast reduces the chance of (significant) stockout and overstock when the reality doesn’t match the most likely scenario, blended should be the first choice. Especially since it’s better to be a few units short than have 10,000 extra or a few months of extra inventory that will eventually sell then be without for three months (due to long lead times).

In fact, the best method is blended models guided by human judgement that can identify the most likely scenarios, validate the best models, and then let the raw computational power of todays platform compute a solution that is near optimal across all of the different forecasts, rather than over optimizing for the wrong forecast. When you don’t know what the actual demand will be, it’s better to go with a blended solution that might be 2% off of optimal than pick the wrong scenario and end up with an “optimized” forecast that is 20% off.

Despite all of our advances, forecasting is still as much art as science, and only expert human judgement on blended models will produce sound forecasts in today’s rocky economic environment.

Where’s My Stuff?

To this day, most C-Suites are still asking Procurement to Show Me The Money! (as they’ve failed to learn anything over the past two decades) when the only question they should be asking is Where’s My Stuff?

Globalization is over. (Well, it’s not over, but it’s not going to be anything like it was for a long, long, time, if ever.)

Free trade is a thing of the past and will soon be regulated to myth.

The steady, five-fold, increase of natural disasters (as well as their unpredictable rapid onslaught) over the last 50 years make any route as dangerous as it has been in centuries, nullifying any improvements we’ve made in weather prediction or transport technology.

Even if you find a calm ocean or air route, the rise in piracy and war still put your vessels at risk.

This is assuming you can even get your goods onto a boat or plane as geopolitical tensions can result in border closings or trade embargoes at any time, escalating worker strife is resulting in strikes that shutdown ports (and airports) for extended periods of time, and man-made disasters (as well as terrorism) can result in your source of supply being cut off completely.

Money is important, because with rampant inflation you have to start counting pennies again, and shave every cent you can from cost, but it’s all irrelevant if you can’t get your stuff and have nothing to sell.

Procurement needs to go back to its roots and focus first on supply assurance, not savings opportunities. Risk, quality, delivery method, and lead time can’t be an afterthought — they need to be the primary up-front focus!

Product designs need to be vetted for realism. The perfect, or even preferred, product design is worthless if it relies on a component that can only be single sourced, (too much of) a raw material or rare earth that is in short supply (and where the majority of supply comes from a single country where supply could be cut off at any time), or the supply chain that would be required increases organizational supply chain complexity.

Suppliers need to be fully evaluated not just for their ability to supply a product that meets the specs, but all of the risks based on who they are, where they are located, the fulfillment routes available to get the product from them, as well as the suppliers and fulfillment routes they rely on for their critical raw materials and parts. A 360-degree risk profile is mandatory.

Carriers need to be evaluated against their own risk profiles which must include financial status, access to affordable fuel, primary fleet location(s), workforce availability, and other factors you may have previously ignored.

Routes need to evaluated not just for delivery time and cost, but safety and reliability with respect to the risk of natural disasters, the current risk of piracy and war, and the reliance on straits (which are all risky due to war and droughts).

Internal operations need to be analyzed for dependencies. Are you using production lines, control systems, or overly customized and unique software solutions that can only be supported by a single vendor which, in unstable economies, can suddenly find themselves bankrupt.

Labour requirements have to be analyzed as well — key roles that encapsulate key knowledge required to organizational success should not be contracted out or handed over to third parties, especially when dependency can mean organizational death!

Only when the risks have been mitigated and the likelihood of supply and continued operation assured do you move on to the money management.

After all, Finance shouldn’t be asking you to Show Me The Money when it’s meaningless unless you have a product that will make the customers Show You The Money.

AI is Not Just New Fangled Technology, It’s Unneeded Technology!

This classic guest post from sixteen years ago is just as relevant today as it was then. Especially since the new fangled technology hallucinates more than we ever could!

 

I say, what is with this new fangled procurement technology?

Back in my day we used a ledger
and a calculator
and we were able to track our costs just fine.
When it came time to negotiations, we offered incentive carrots
like free meals at fine upscale restaurants
and free trips to exotic locales like Aruba and Cancun.
If that didn’t work we brought out the big stick
and threatened to take our business to the supplier down the street
or to expose all the kickbacks the supplier rep took from the last deal.
We didn’t need any new fangled e-Sourcing technology to conduct our negotiations
or marketplace technology to handle our invoices
or e-Payment systems to pay our bills.
When we received the goods at our warehouse
we just handed over a big wad of cash
and everyone was happy.
Now get off my lawn!

Two By Two They’re Still Going Down …

… because the satellite circus won’t leave town.

And continues to push their 2*2s as your solution to all your tech woes. It’s not just Gartner and their Magic Quadrant, although that is usually the first map pushed in front of C-Suite Executives as part of organized corporate brainwashing — usually by Big Consultancies and vendors who sponsor the map in order to push it in front of their C-Suite targets.

Especially when the vendor, or the vendor that the consultancy has diamond/platinum/sycophant status with, is in the upper right quadrant which — according to map definition — makes the vendor one of the de-factor best vendors, no matter what.

And it’s complete and utter bullcrap exceeded only by the Gen-AI bullcrap.

So when an executive tells you they need to use vendor X because the Big Analyst 2 * 2 says so, you need to have a systematic deconstruction playbook at your disposal. In fact, you need to have it memorized because they will spring it on you just as they are about to sign a contract and you will have to stop them dead in their tracks.

Which means you need to understand everything about what the 2 * 2 maps aren’t.

They aren’t best tech, or even best vendor.

They aren’t exhaustive, or even modern.

They aren’t unbiased, or even consistent.

In a post from this summer, which probably flew under your radar, Joël Collin-Demers tackled busting the Gartner MQ because, as the most popular map, it’s the one you are most likely going to have to refute.

In his post, Joël makes it clear that:

“Leader” status measures market presence (and the analyst firm’s definition of vision, which usually correlates to how many zeroes they can see on the cheque the vendor writes)

“Transformation” capability is not measured, nor is real client success.

“Strategy” measured is theirs, not yours.

And many other things.

Moreover, it ignores the fact that:

“Subjectivity” is the name of the game. There’s no well defined scale because the majority of factors being measured are purely subjective.

“Axes are Meaningless” as they typically squish half-a-dozen factors into one, whether they fit together or not.

“Different is same” because six different factors averaged into one means that completely different company profiles become indistinguishable if the average is the same.

In other words, while the maps can point you to some vendors that are, or at least were, rather significant in the space that you might want to consider adding to your potential vendor list that you are considering for your RFP, that’s all these maps are good for. Creating a partial starting list. No more, no less. Especially since, as you do more research and hone in on your industry, strategy, spend profile, actual needs, and financial means, they might all fall off that list as you identify vendors that are likely to be better for your needs and invite them to the RFP instead.

So be sure to check out Joël Collin-Demers post where he tackled Gartner MQ Busting and start creating a playbook you can immediately bring to bear on a moment’s notice.