About Jeff Harrop

Employing the Law of Large Numbers in Bottom-Up Forecasting

Posted on June 5, 2019 by Jeff Harrop

It is utterly implausible that a mathematical formula should make the future known to us, and those who think it can would once have believed in witchcraft. – Jakob Bernoulli (1655-1705)

This is a topic I’ve touched on numerous times in the past, but I’ve never really taken the time to tackle the subject comprehensively.

Before diving in, I just want to make clear that I’m going to stay in my lane: the frame of reference for this entire piece is around forecasting sales at the point of consumption in retail.

In that context, here are some truths that I consider to be self evident:

Consumers buy specific items in specific stores at specific times. Therefore, in order to plan the retail supply chain from consumer demand back, forecasts are needed by item by store.
Any retailer has a large enough percentage of intermittent demand streams at item/store level (e.g. fewer than 1 sale per week) that they can’t simply be ignored in the forecasting process.
Any given item can have continuous demand in some locations and intermittent demand in other locations.
“Intermittent” doesn’t mean the same thing as “random”. An intermittent demand stream could very well have a distinct pattern that is not visible to the naked eye (nor to most forecast algorithms that were designed to work with continuous demands).
Because of points 1 to 4 above, the Law of Large Numbers needs to be employed to see any patterns that exist in intermittent demand streams.

On this basis, it seems to be a foregone conclusion that the only way to forecast at item/store is by employing a top-down approach (i.e. aggregate sales history to some higher level(s) than item/store so that a pattern emerges, calculate an independent forecast at that level, then push down the results proportionally to the item/stores that participated in the original aggregation of history).

So now the question becomes: How do you pick the right aggregation level for forecasting?

This recent (and conveniently titled) article from Institute of Business Forecasting by Eric Wilson called How Do You Pick the Right Aggregation Level for Forecasting? captures the considerations and drawbacks quite nicely and provides an excellent framework to discuss the problem in a retail context.

A key excerpt from that article is below (I recommend that you read the whole thing – it’s very succinct and captures the essence about how to think about this problem in a short few paragraphs):

When To Go High Or Low?

Despite all the potential attributes, levels of aggregation, and combinations of them, historically the debate has been condensed down to only two options, top down and bottom up.

The top-down approach uses an aggregate of the data at the highest level to develop a summary forecast, which is then allocated to individual items on the basis of their historical relativity to the aggregate. This can be any generated forecast as a ratio of their contribution to the sum of the aggregate or on history which is in essence a naïve forecast.

More aggregated data is inherently less noisy than low-level data because noise cancels itself out in the process of aggregation. But while forecasting only at higher levels may be easier and provides less error, it can degrade forecast quality because patterns in low level data may be lost. High level works best when behavior of low-level items is highly correlated and the relationship between them is stable. Low level tends to work best when behavior of the data series is very different from each other (i.e. independent) and the method you use is good at picking up these patterns.

The major challenge is that the required level of aggregation to get meaningful statistical information may not match the precision required by the business. You may also find that the requirements of the business may not need a level of granularity (i.e. Customer for production purposes) but certain customers may behave differently, or input is at the item/customer or lower level. More often than not it is a combination of these and you need multiple levels of aggregation and multiple levels of inputs along with varying degrees of noise and signals.

These are the two most important points:

“High level works best when behavior of low-level items is highly correlated and the relationship between them is stable.”
“Low level tends to work best when behavior of the data series is very different from each other (i.e. independent) and the method you use is good at picking up these patterns.”

Now, here’s the conundrum in retail:

The behaviour of low level items is very often NOT highly correlated, making forecasting at higher levels a dubious proposition.
Most popular forecasting methods only work well with continuous demand history data, which can often be scarce at item/store level (i.e. they’re not “good at picking up these patterns”).

My understanding of this issue was firmly cemented about 19 years ago when I was involved in a supply chain planning simulation for beer sales at 8 convenience stores in the greater Montreal area. During that exercise, we discovered that 7 of those 8 stores had a sales pattern that one would expect for beer consumption in Canada (repeated over 2 full years): strong sales during the summer months, lower sales in the cooler months and a spike around the holidays. The actual data is long gone, but for those 7 stores, it looked something like this:

The 8th store had a somewhat different pattern.

And by “somewhat different”, I mean exactly the opposite:

Remember, these stores were all located within about 30 kilometres of each other, so they all experienced generally the same weather and temperature at the same time. We fretted over this problem for awhile, thinking that it might be an issue with the data. We even went so far as to call the owner of the 8 store chain to ask him what might be going on.

In an exasperated tone that is typical of many French Canadians, he impatiently told us that of course that particular store has slower beer sales in the summer… because it is located in the middle of 3 downtown university campuses: fewer students in the summer months = a decrease in sales for beer during that time for that particular store.

If we had visited every one of those 8 stores before we started the analysis (we didn’t), we may have indeed noticed the proximity of university campuses to one particular store. Would we have pieced together the cause/effect relationship to beer sales? My guess is probably not. Yet the whole story was right there in the sales data itself, as plain as the nose on your face.

We happened upon this quirk after studying a couple dozen SKUs across 8 locations. A decent sized retailer can sell tens of thousands of SKUs across hundreds or thousands of locations. With millions of item/store combinations, how many other quirky criteria like that could be lurking beneath the surface and driving the sales pattern for any particular item at any particular location?

My primary conclusion from that exercise was that aggregating sales across store locations is definitely NOT a good idea.

So in terms of figuring out the right level of aggregation, that just leaves us with the item dimension – stay at store level, but aggregate across categories of similar items. But in order for this to be a good option for the top level, we now have another problem: “behavior of low-level items is highly correlated and the relationship between them is stable“.

That second part becomes a real issue when it comes to trying to aggregate across items. Retailers live every day on the front line of changing consumer sentiment and behaviour. As a consequence of that, it is very uncommon to see a stable assortment of items in every store year in and year out.

Let’s say that a category currently has 10 similar items in it. After an assortment review, it’s decided that 2 of those items will be leaving the category and 4 new products will be introduced into the category. This change is planned to be executed in 3 months’ time. This is a very simple variation of a common scenario in retail.

Now think about what that means with regard to managing the aggregated sales history for the top level (category/store):

The item/store sales history currently includes 2 items that will be leaving the assortment. But you can’t simply exclude those 2 items from the history aggregation, because this would understate the category/store forecast for the next 3 months, during which time those 2 items will still be selling.
The item/store level sales history currently does not include the 4 new items that will be entering the assortment. But you can’t simply add surrogate history for the 4 new items into the aggregation, because this would overstate the category/store forecast for next 3 months before those items are officially launched.

In this scenario, how would one go about setting up the category/store forecast in such a way that:

It accounts for the specific items participating in the aggregation at different future times (before, during and after the anticipated assortment change)?
The category/store forecast is being pushed down to the correct items at different future times (before, during and after the anticipated assortment change)?

And this is a fairly simple example. What if the assortment changes above are being rolled out to different stores at different times (e.g. a test market launch followed by a staged rollout)? What if not every store is carrying the full 10 SKU assortment today? What if not every store will be carrying the full 12 SKU assortment in the future?

The complexity of trying to deal with this in a top-down structure can be nauseating.

So it seems that we find ourselves in a bit of a pickle here:

The top-down approach is unworkable in retail because the behaviour between locations for the same item are not correlated (beer in Montreal stores) and the relationships among items for the same location are not stable (constantly changing assortments).
In order for the bottom-up approach to work, there needs to be some way of finding patterns in intermittent data. It’s a self-evident truth that the only way to do this is by aggregating.

So the Law of Large Numbers is still needed to solve this problem, but in a retail setting, there is no “right level” of aggregation above item/store at which to develop reliable independent top level forecasts that are also manageable.

Maybe we haven’t been thinking about this problem in the right way.

This is where Darryl Landvater comes in. He’s a long time colleague and mentor of mine best known as a “manufacturing guy” (he’s the author of World Class Production and Inventory Management, as well as co-author of The MRP II Standard System), but in reality he’s actually a “planning guy”.

A number of years ago, Darryl recognized the inherent flaws with using a top-down approach to apply patterns to intermittent demand streams and broke the problem down into two discrete parts:

What is the height of the curve (i.e. rate of sale)?
What is the shape of the curve (i.e. selling profile)?

His contention was that it’s not necessary to use aggregation to calculate completely independent sales forecasts (i.e. height + shape) to achieve this. Instead, what’s needed is to aggregate to calculate selling profiles to be used in cases where the discrete demand history for an item at a store is insufficient to determine one. We’re still using the Law of Large Numbers, but only to solve for the specific problem inherent in slow selling demands – finding the shape of the curve.

It’s called Profile Based Forecasting and here’s a very simplified explanation of how it works:

Calculate an annual forecast quantity for each independent item/store based on sales history from the last 52+ weeks (at least 104 weeks of rolling history is ideal). For example, if an item in a store sold 25 units 2 years ago and 30 units over the most current 52 weeks, then the total forecast for the upcoming 52 weeks might be around 36 units with a calculated trend applied.
Spread the annual forecast into individual time periods as follows:
- If the item/store has a sufficiently high rate of sale that a pattern can be discerned from its own unique sales history (for example, at least 70 units per year), then calculate the selling pattern from only that history and multiply it through the item/store’s selling rate.
- If the item/store’s rate of sale is below the “fast enough to use its own history” threshold, then calculate a sales pattern using a category of similar items at the same store and multiply those percentages through the independently calculated item/store annual forecast.

There is far more to it than that, but the separation of “height of the curve” from “shape of the curve” as described above is the critical design element that forms the foundation of the approach.

Think about what that means:

If an item/store’s rate of sale is sufficient to calculate its own independent sales profile at that level, then it will do so.
If the rate of sale is too low to discern a pattern, then the shape being applied to the independent item/store’s rate of sale is derived by looking at similar items in the category within the same store. Because the profiles are calculated from similar products and only represent the weekly percentages through which to multiply the independent rate of sale, they don’t need to be recalculated very often and are generally immune to the “ins and outs” of specific products in the category. It’s just a shape, remember.
All forecasting is purely bottom-up. Every item at every store can have its own independent forecast with a realistic selling pattern and there are no forecasts to be calculated or managed above the item/store level.
The same forecast method can be used for every item at every store. The only difference between fast and slow selling items is how the selling profile is determined. As the selling rate trends up or down over time, the appropriate selling profile will be automatically applied based on a comparison to the threshold. This makes the approach very “low touch” – demand planners can easily oversee several hundred thousand item/store combinations by managing only exceptions.

With realistic, properly shaped forecasts for every item/store enabled without any aggregate level modelling, it’s now possible to do top-down stuff that makes sense, such as applying promotional lifts or overrides for an item across a group of stores and applying the result proportionally based on each store’s individual height and shape for those specific weeks, rather than using a naive “flat line” method.

Simple. Intuitive. Practical. Consistent. Manageable. Proven.

A Digitally Connected, Consumer-Driven Supply Chain

Posted on May 14, 2019 by Jeff Harrop

Check out this article about Princess Auto’s Flowcasting journey from Supply Chain Management Review, authored by Mike Doherty and Sylvain Landry.

Also available on our Resources page.

Customer Service Collateral Damage

Posted on April 3, 2019 by Jeff Harrop

Good intentions can often lead to unintended consequences. – Tim Walberg

Speed kills.

Retailers with brick and mortar operations are always trying to keep the checkout lines moving and get customers out the door as quickly as possible. Many collect time stamps on their sales transactions in order to measure and reward their cashiers based on how quickly they can scan.

Similarly, being able to receive quickly at the back of the store is seen as critical to customer service – product only sells off the shelf, not from the receiving bay or the back of a truck.

This focus on speed has led to many in-store transactional “efficiencies”:

If a customer puts 12 cans of frozen concentrated juice on the belt, a cashier may scan the first one and use the multiplier key to add the other 11 to the bill all at once.
If a product doesn’t scan properly or is missing the UPC code, just ask the customer for the price and key the sale under a “miscellaneous” SKU or a similar item with the same price, rather than calling for a time consuming code check.
If a shipment arrives in the receiving bay, just scan the waybill instead of each individual case and get the product to the floor.

These time saving measures can certainly delight “the customer of this moment”, but there can also be consequences.

In the “mult key” example, the 12 cans scanned could be across 6 different flavours of juice. The customer may not care since they’re paying the same price, but the inventory records for 6 different SKUs have just been fouled up for the sake of saving a few seconds. To the extent that the system on hand balances are used to make automated replenishment decisions, this one action could be inconveniencing countless customers for several more days, weeks or even months before the lie is exposed.

The smile on a customer’s face because you saved her 5 seconds at the checkout or the cashier speed rankings board in the break room might be tangible signs of “great customer service”, but the not-so-easy-to-see costs of stockouts and lost sales that arise from this practice over time is extremely costly.

Similarly with skipping code checks or “pencil whipping” back door receipts. Is sacrificing accuracy for the sake of speed really good customer service policy?

A recent article published in Canadian Grocer magazine begins with the following sentence:

“A lack of open checkouts and crowded aisles may be annoying to grocery shoppers, but their biggest frustration is finding a desired product is out of stock, according to new research from Field Agent.”

According to the article, out of stocks are costing Canadian grocers $63 billion per year in sales. While better store level planning and replenishment can drive system reported in-stocks close to 100%, the benefits are muted if the replenishment system thinks the store has 5 units when they actually have none.

Not only does this affect the experience of a walk-in customer looking at an empty shelf, but it’s actually even more serious in an omnichannel world where the expectation is that retailers will publish store inventories on their public websites (gulp!). An empty shelf is one thing, but publishing an inaccurate on hand on your website is tantamount to lying right to your customers’ faces.

We’ve seen firsthand that it’s not uncommon for retailers to have a store on hand accuracy percentage in the low 60s (meaning that almost 40% of the time, the system on hand record differs from the counted quantity by more than 5% at item/location level). Furthermore, we’ve found that on the day of an inventory count, the actual in stock is several points lower than the reported in stock on average.

Suffice it to say that inaccurate on hand records are a big part of the out of stocks problem.

Nothing I’ve said above is particularly revolutionary or insightful. The real question is why has it been allowed to continue?

In my view, there are 3 key reasons:

Most retailers conflate shrink with inventory accuracy and make the horribly, horribly mistaken assumption that if their financial shrink is below 1.5%, then their inventory management is under control. Shrink is a measure for accountants, not customers and the responsibility of store inventory management belongs in Store Operations, not Finance.
Nobody measures the accuracy of their on hands. It’s fine to measure the speed of transactions and the efficiency of store labour, but if you’re taking shortcuts to achieve those efficiencies, you should also be measuring the consequence of those actions – especially when the consequence so profoundly impacts the customer experience.
Retailers think that inaccurate store on hands is an intractable problem that’s impossible to economically solve. That was true for every identified problem in human history at one point. However, I do agree that if no action is taken to solve the problem because it is “impossible to solve”, then it will never be solved.

It’s true that overcoming inertia on this will not be easy.

Your customers’ expectations will continue to rise regardless.

Rise of the Machines?

Posted on February 6, 2019 by Jeff Harrop

It requires a very unusual mind to undertake the analysis of the obvious. – Alfred North Whitehead (1861-1947)

My doctor told me that I need to reduce the amount of salt, fat and sugar in my diet. So I immediately increased the frequency of oil changes for my car.

Confused?

I don’t blame you. That’s how I felt after I read a recent survey about the adoption of artificial intelligence (AI) in retail.

Note that I’m not criticizing the survey itself. It’s a summary of collected thoughts and opinions of retail C-level executives (pretty evenly split among hardlines/softlines/grocery on the format dimension and large/medium/small on the size dimension), so by definition it can’t be “wrong”. I just found some of the responses to be revealing – and bewildering.

On the “makes sense” side of the ledger, the retail executives surveyed intend to significantly expand customer delivery options for purchases made online over the next 24 months, specifically:

79% plan to offer ship from store
80% plan to offer pick up in store
75% plan to offer delivery using third party services

This supports my (not particularly original) view that the physical store affords traditional brick and mortar retailers a competitive advantage over online retailers like Amazon, at least in the short to medium term.

However, the next part of the survey is where we start to see trouble (the title of this section is “Retailers Everywhere Aren’t Ready for the Anywhere Shelf”):

55% of retailers surveyed don’t have a single view of inventory across channels
78% of retailers surveyed don’t have a real-time view of inventory across channels

What’s worse is that there is no mention at all about inventory accuracy. I submit that the other 45% and 22% respectively may have inventory visibility capabilities, but are they certain that their store level inventory records are accurate? Do they actually measure store on hand accuracy (by item by location in units, which is what a customer sees) as a KPI?

The title of the next slide is “Customer Experience and Supply Chain Maturity Demands Edge Technologies”. Okay… Sure… I guess.

The slide after that concludes that retail C-suite executives believe that the top technologies “having the broadest business impact on productivity, operational efficiency and customer experience” are as follows:

#1 – Artificial Intelligence/Machine Learning
#2 – Connected Devices
#3 – Voice Recognition

Towards the end, it was revealed that “The C-suite is planning a 5X increase in artificial intelligence adoption over the next 2 years”. And that 50% of those executives see AI as an emerging technology that will have a significant impact on “sharpening inventory levels” (whatever that actually means).

So just to recap:

Over the next 2 years, retailers will be aggressively pursuing customer delivery options that place ever increasing importance on visibility and accuracy of store inventory.
A majority of retailers haven’t even met the visibility criteria and it’s highly unlikely that the ones who have are meeting the accuracy criteria (the second part is my assumption and I welcome being proved wrong on that).
Over the next 2 years, retailers intend to increase their investment in artificial intelligence technologies fivefold.

I’m reminded of the scene in Die Hard 2 (careful before you click – the language is not suitable for a work environment or if small children are nearby) where terrorists take over Dulles International Airport during a zero visibility snowstorm and crash a passenger jet simply by transmitting a false altitude reading to the cockpit of the plane.

Even in 1990, passenger aircraft were quite technologically advanced and loaded with systems that could meet the definition of “artificial intelligence“. What happens when one piece of critical data fed into the system is wrong? Catastrophe.

I need some help understanding the thought process here. How exactly will AI solve the inventory visibility/accuracy problem? Are we talking about every retailer having shelf scanning robots running around in every store 2 years from now? What does “sharpen inventory levels” mean and how is AI expected to achieve that (very nebulous sounding) goal?

I’m seriously asking.

Concealing Your Shame

Posted on December 5, 2018 by Jeff Harrop

There is no shame in not knowing; the shame lies in not finding out. – Russian Proverb

Customer expectations of brick & mortar retailers are changing.

Most retailers are failing miserably at meeting those expectations with regard to providing information about stock availability at their stores online.

I’m not talking about whether or not they have sufficient stock to meet customer demand – it’s even more basic than that. When a customer is looking to visit your store can you even properly tell him/her what your stock status actually is?

Recently, I decided to anecdotally put one particular store to the test on this. I chose this store for the following reasons:

They actually publish their store on hand balances online for all the world to see in real time.
They offer a “buy online, pick up in store” option.
I visit the store fairly frequently and it’s about 1 kilometre from my house.

On the day of my “study”, I only had 2 items I needed. Before leaving, I called up the pages for those items on my iPhone and went to the store. When I got there, I refreshed the pages to retrieve the most up-to-date stock information and compared that number to what I actually found on the shelf. After that, I wandered around the aisles and picked a few other items at random and did the same thing.

Now before I share the results, there are some rather significant caveats that I need to mention:

The inventory is updated in real time, but obviously it’s based on POS transactions. When I did the “physical count” on the shelf, it’s certainly possible that some other customer had picked the item off the shelf but had not yet paid for it.
The study was performed on a busy Saturday afternoon about 4 weeks before Christmas. Not exactly ideal timing for ensuring that the store was stocked neatly or that there wasn’t a lot of product floating around in customer baskets as per point 1 above.
I know that this store has a very large back room and doesn’t keep separate on hand balances for shelf stock and backroom stock. In cases where my count is short, it’s certainly possible that the product was in the back room or displayed elsewhere in the store.
When I got a count discrepancy, I did not ask the staff for help in locating the “missing” items. As I mentioned, we are only weeks away from Christmas and I wasn’t about to waste people’s time finding items that I had no intention of purchasing.

The first item on my list was a carbon dioxide cylinder for our SodaStream. Note that I’ve attempted to crop out any information that would reveal who the retailer is (logos, shelf tags, product identifiers, etc.). This won’t stop some of you from recognizing them, but I can’t do much about that.

Okay, back to the SodaStream cylinder. When I reached the shelf and refreshed the page on my phone, here’s what I got:

Wow, 337 units in stock! (As an aside, this retailer almost always shows the aisle number in the store where the product can be found, which is stellar – not sure why it’s not shown in this case, but it’s a product I buy often, so I knew exactly where to go).

Now here’s the shelf:

You can’t see them all in this image, but the actual count was 18 units, far short of 337. Obviously this is either a massive inventory record error or there’s a pallet of them on a secondary display or in the back room. So long as they sell fewer than 18 per day, buyers of this item will be happy.

RESULT: INCONCLUSIVE

The second item on my list was a large, bark deterring dog collar for my mother-in-law’s dog (it uses vibration or noise to deter barking, not electric shocks, so don’t judge me!). As you’ll see below, my phone told me to go to aisle 56 to find 1 unit:

Unfortunately when I got to the aisle, there was none to be found. I spent a few minutes searching all of the overheads, pegs and bins in this aisle and one aisle over in each direction and couldn’t find it.

RESULT: FAIL

While in aisle 56, I picked another random item (mulberry scented dog shampoo) and looked it up on my phone:

And here is the shelf:

6 units – right on the nose.

RESULT: SUCCESS

Now, how about this Bissell Little Green pet stain remover?

This item is on promotion for $25.00 off and I found an end aisle display with 12 units:

…and one more unit in the home in aisle 60:

So that’s 13 on the shelf vs 32 units reported on hand. But because this item is promoted, there is almost certainly more in the back room to replenish the shelves.

RESULT: INCONCLUSIVE

On to aisle 17 to check out the Stanley chalk line reels.

Hoping to find 5…

…and 5 it is.

RESULT: SUCCESS

You get the picture (no pun intended). I also documented a few other items in the same way, but I’ll spare you the photographic evidence:

Richard Self Adhesive Drywall Tape: 3 online, 4 on the shelf (RESULT: PRETTY CLOSE)
T.S.P. Heavy Duty Cleaner (400g): 10 online, 4 on the shelf (RESULT: FAIL)
Soft Glide Cabinet Hinge: 12 online, none to be found anywhere (RESULT: EPIC FAIL)
OOK Picture Hanging Kit: 14 online, 13 on the peg (RESULT: PRETTY CLOSE)

In summary:

There were 3 failures out of 9 (I’m counting “Pretty Close” and “Inconclusive” in the success column for fairness)
2 of those 3 failures could have resulted in a lost sale on that day (i.e. the reported on hand was > 0, but there was no stock to be found on the sales floor).
With regard to the bark deterrent collar (one of the items I actually wanted to buy), there’s more to the story:
- When I got home, I ordered the item for in store pickup and the on hand immediately dropped to zero
- Later that day, I received an email notification and a phone call informing me that the item wouldn’t be available for pickup until the next day
- From this, I’m surmising that they couldn’t find it in the store and had one delivered from a nearby store overnight
- The next day, I picked up the item at my home store – lost sale averted

So what was the point of all this and why did I choose “Concealing Your Shame” as the title? Am I trying to shame this retailer for what (anecdotally and with all of my previous caveats applied) looks like imperfect performance?

Au contraire!

Store on hand accuracy is not easy to achieve and this retailer is to be highly commended for their confidence and willingness to be as transparent to customers as possible.

No, the shame is reserved for those retailers who have on hand balances readily available in their systems but choose not to share it. I guess the thinking is that you can’t fail if you don’t try.

I say it again: customer expectations are changing.

If you’re afraid to share your on hand balances with your customers, I have 2 questions:

Why? (you already know why)
What are you doing about it?

Accuracy or Precision?

Posted on October 3, 2018 by Jeff Harrop

It is the mark of an educated mind to rest satisfied with the degree of precision which the nature of the subject admits and not to seek exactness where only an approximation is possible. – Aristotle (384 BC – 322 BC)

My favourite part about writing these articles is finding just the right quote to introduce them. Before we get started, go back and read the quote from Aristotle above if you happened to skip past it – I think it both accurately and precisely summarizes my argument.

Now in the context of forecasting for the supply chain, let’s talk about what each of these terms mean:

Accuracy: Ability to hit the target (i.e. how close is the actual to the forecast?)

Precision: Size of the target you’re aiming at (i.e. specificity of product, place and timing of the forecast)

I’m sorry to be a total downer, but the reason this article is titled Accuracy or Precision is because you can’t have both. The upper right quadrant in the illustration above ain’t happening (a bit more on that later).

In the world of forecasting, people seem obsessed with accuracy and often ask questions like:

What level of forecast accuracy are you achieving?
How should we be benchmarking our forecast accuracy?
Are we accurate enough? How can we be more accurate?

The problem here is that any discussion about forecast accuracy that does not at the same time account for precision is a complete waste of time.

For example, one tried and true method for increasing forecast accuracy is by harnessing the mystical properties of The Law of Large Numbers.

To put it another way – by sacrificing precision.

Or to put it in the most cheeky way possible (many thanks to Richard Sherman for this gem, which I quote often):

Sherman’s Law:
Forecast accuracy improves in direct correlation to its distance from usefulness.

So how do we manage the tradeoff between precision and accuracy in forecasting?You must choose the level of precision that is required (and no more precise than that) and accept that in doing so, you may be sacrificing accuracy.

For a retailer, the only demand that is truly independent is customer demand at the point of sale. Customers choose specific items in specific locations on specific days. That’s how the retail business works.

This means that the precision of the forecasting process must be by item by location by day – full stop.

Would you be able to make a more accurate prediction by forecasting in aggregate for an item (or a group of items) across all locations by month? Without a doubt.

Will that help you figure out when you need to replenish stock for a 4 pack of 9.5 watt A19 LED light bulbs at store #1378 in Wichita, Kansas?

Nope. Useless.

I can almost see the wincing and hear the heart palpitations that this declaration will cause.

“Oh God! You’ll NEVER be able to get accurate forecasts at that level of precision!” To that I say two things:

It depends on what level of accuracy is actually required at that level of precision.
Too damn bad. That’s the requirement as per your customers’ expectation.

With regard to the first point, keep in mind that it’s not uncommon for an item in a retail store to sell fewer than 20 units per YEAR. On top of that, there are minimum display quantities and pack rounding that will ultimately dictate how much inventory will be available to customers to a much greater degree than the forecast.Forecasts by item/location/day are still necessary to plan and schedule the upstream supply chain properly, but it’s only necessary for forecasts at that level of precision to be reasonable, not accurate in the traditional sense of the word. This is especially true if you also replan daily with refreshed sales and inventory numbers for every item at every location.

There are those out there who would argue that my entire premise is flawed. That I’m not considering the fact that with advances in artificial intelligence, big data and machine learning, it will actually be possible to process trillions of data elements simultaneously to achieve both precision and accuracy. That I shouldn’t even be constraining my thinking to daily forecasting – soon, we’ll be able to forecast hourly.

Let’s go back to the example I mentioned earlier – an item that sells 20 units (give or take) in a location throughout the course of a year. Assuming that store is open for business 12 hours out of every day and closed 5 days per year for holidays, there are 4,320 hours in which those 20 units will sell. Are we to believe that collecting tons of noise (whoops, I meant “data”) from social media, weather forecasting services and the hourly movement of soybean prices (I mean, why not, right?) will actually be able to predict with accuracy the precise hour for each of those 20 units in that location over the next year? Out of 4,320 hours to choose from? Really?

(Let’s put aside the fact that no retailer that I’ve ever seen even measures how accurate their on hand records are right now, let alone thinking they can predict sales by hour).

I sometimes have a tendency to walk the middle line on these types of predictions. “I don’t see it happening anytime soon, but who knows? Maybe someday…”

Well, not this time.

This is utter BS. Unless all of the laws of statistics have been debunked recently without my noticing, degrees of freedom are still degrees of freedom.

Yes, I’m a loud and proud naysayer on this one and if anyone ever actually implements something like that and demonstrates the benefits they’re pitching, I will gleefully eat a wheelbarrow of live crickets when that time comes (assuming I’m not long dead).

In the meantime, I’m willing to bet my flying car, my personal jetpack and my timeshare on the moon colony (all of which were supposed to be ubiquitous by now) that this will eventually be exposed as total nonsense.

Customer Convenience in the Eye of the Beholder

Posted on August 1, 2018 by Jeff Harrop

Beauty is in the eye of the beholder and it may be necessary from time to time to give a stupid or misinformed beholder a black eye. – Miss Piggy

From a customer’s point of view, what is convenience?

Let’s start with the basic premise that customers are looking for:

Products they want
Where they want it
At a price they’re willing to pay

Thankfully, the rhetoric around the so-called “retail apocalypse” has slowly morphed from “Everyone except Amazon is dead!” to “Retailers who are doing poorly at all 3 of those things are dead.”

To be sure, the era of online retailing ushered in by Amazon has moved the goal posts with regard to all 3 of those customer criteria. Traditional brick and mortar retailers have picked up the mantle and have actually (in my opinion) put Amazon at a bit of a disadvantage (for now) by offering customers the option to buy products online and pick them up in a nearby store.

But the experiences customers are looking for vary not only by customer, but also by the nature and reason for the transaction. In the last month, I have purchased products using the 3 most common delivery methods (buy online with home delivery, buy online and pick up in store, and kicking it old school by walking into a store, filling up the cart and checking out).

After reflecting on each of those experiences, I came to realize that, as a customer, I need different types of “convenience” in different situations and every shopping experience comes with positives and negatives.

Scenario 1: Buy Online with Home Delivery

The leather satchel that I’ve been using for years has started to get a little long in the tooth (2 zippers were broken and the shoulder strap was nearly ready to snap). As a consultant, I travel frequently with my laptop (along with all the accessories that go with it) and I thought that I would switch things up and go with a padded backpack instead.

I had a very specific set of needs and I wasn’t in a huge rush to get it. However, I envisioned a scenario whereby I would have to visit multiple stores in order to find what I was looking for at a reasonable price. In this case (for me), Amazon was the answer.

I began with a search for “laptop backpacks”, got approximately 17 trillion results and started narrowing things down from there. Truth be told, the number of choices made shopping a less than pleasurable experience for me, as it took a fair amount of time to read descriptions, dimensions, features, etc. before I was able to come to a decision. But I feel that getting the exact product that I wanted from all available choices was worth the time and inconvenience of having to wait a couple days to get the product, while also not having to run all over creation hoping to find it on a shelf somewhere.

Scenario 2: Kickin’ It Old School

Okay, I admit it – I enjoy grocery shopping. I always go with a list and I always come home with a few items that weren’t on my list. I also come home without items that were on my list because the store was out of stock.

I’m not really much of an impulse shopper. More often than not, when I buy an item that’s not on my list, I do so because seeing it on the shelf is a reminder that I actually need it, but I neglected to put it on my list.

Thus far, online shopping has been woefully inadequate at replicating that experience, which is what leads me to believe that the traditional supermarket is here to stay for quite some time yet. I couldn’t imagine putting together an online grocery order for 50+ items.

That said, it would be nice if they were in stock, which leads me to my next experience…

Scenario 3: Buy Online, Pick Up In Store

It was Wednesday before an annual weekend (25 years running by my count) where a bunch of us head up to a friend’s cottage for 3 days of music, laughs, fairly heavy drinking and general merriment. The weekend traditionally begins on Thursday night with a steak barbecue.

On Tuesday of that week, I noticed that my local grocery store had frozen lobster tails on sale for $3.50 each, so I thought I’d grab a dozen of them to accompany the steaks on Thursday night. I had a few other essentials to pick up while I was there, so I composed a small list and went to the store. I went straight to the seafood department first and they were out of lobster tails. That was my main reason for going, so I left without getting anything else on the list.

That same grocer had another store not far away, but I figured my chances of getting lobster tails there was low, plus I didn’t have the additional time to spend on a treasure hunt that may turn up nothing. Disappointed, I went home.

As I was doing the normal work from home stuff (documents, emails, conference calls, etc.), I remembered that this store offered a buy online, pick up in store service, so I navigated to their website and placed an order for 12 lobster tails (plus the 6 or so items that were on my list) to be picked up the next day (Wednesday), which would still give me time to thaw them in the refrigerator overnight before heading to the cottage on Thursday.

I was fully expecting to be informed by email on Wednesday morning that I was still out of luck on the lobster tails. This would have sucked, but it would have sucked even more if I had wasted another trip to the store in order to find that out.

As luck would have it, I was able to get everything on my list (including the lobster tails) delivered right to my car in the parking lot in an appointment window of my choosing. And yes, the lobster tails were delicious.

Even though that customer experience had a satisfactory ending, was I actually using the BOPIS option for “convenience” or as a stock reservation mechanism due to lack of trust that the store was going to be in stock prior to my arrival?

The point here is that there is no catch-all definition of convenience that suits every customer in every situation. I know people who buy toilet paper online and have it shipped to their house. I also know people who would never buy anything on a website for fear of identity theft. And yes, sometimes trucking your butt to the store and rolling a cart through the aisles is actually the most convenient way to shop.

The truly great retailers of the future will be the ones who can execute well on all 3 of those experiences (plus others that haven’t been invented yet).

Sacred Cows

Posted on June 6, 2018 by Jeff Harrop

I cannot call to mind a single instance where I have ever been irreverant except toward the things which were sacred to other people. – Mark Twain

sacred cow (noun): an idea, custom, or institution held, especially unreasonably, to be above criticism

They are prevalent in all aspects of life and business and the management of the retail supply chain is no exception.

I like to think of sacred cows as “universal assumptions masquerading as universal truths”. What makes them particularly insidious is that there’s a kernel of universal truth in every universal assumption.

Promotional Ordering

The universal truth is that promotions to consumers can have a significant impact on the supply chain known as the “basketball through a garden hose effect”, meaning that operations and suppliers need to see the volume spikes coming well in advance in order to properly prepare for it. Further, promotions are particularly important to retailers, because it’s not a good look when customers are enticed to sacrifice their time and energy to make a trip to the store only to find bare shelves for items that are being promoted.

The universal assumption that springs from this universal truth is that the planning and ordering process for promotions must be separate and distinct from the process for satisfying “baseline” sales. Large orders need to be placed with suppliers months in advance, even though the normal lead time is 7 days. Shipments of promotional goods (both inbound to the DCs and out to the stores) need to be segregated with a big red bow on them so that everyone knows how important they are. And when promotional stock arrives at the DC, it must be immediately allocated to the stores as soon as possible to make sure everyone is ready.

Here’s the problem: Every survey published for the last 20+ years has shown that retail out-of-stocks average 8% overall, but that number climbs to 15% when an item is on promotion. Mission NOT accomplished.

Maybe it’s time to start questioning some of those assumptions:

Do you truly need to violate the universal supply chain principle of postponement by committing to promotional volumes with a longer lead-time? What if you instead just updated your sales forecast and shared a continuously updated time-phased purchasing schedule with suppliers (which they would use for their production planning activities) and only locked in the order when it’s time for them to pick, pack and ship?
Is a “locked in order months in advance” the only way to notify the upstream supply chain about a large need, or can the same time-phased schedule be cubed out to see the impact on operations? Maybe in the grand scheme of everything that the operations will need to deal with during that period (promotional and non promotional volumes), the “big spike” on some particular items coming up on promotion may not even be noticed.
Is moving product to the stores far in advance of a promotion the best way to “be ready” or does it just result in backroom congestion until the promotion begins?

Line Reviews and Planogram Resets

The universal truth about new product launches is that – in a similar vein to promotions – there is a significant volume surge to fill the shelves with new items and considerable uncertainty about what will happen after a new item hits the sales floor. Similarly, the impact of increasing or reducing an item’s space allocation store by store may likewise cause artificial replenishment spikes or troughs.

The universal assumption is that, similar to promotions, the movement of stock from suppliers to stores to support product line changes must have its own unique process – purchase orders with red bows on them, separate allocations of stock to the stores, “DC holdback” schemes, etc.

Some questions spring to mind:

If the Line Review process defines the “start selling date” for new items and the planograms by store for the new product line, why not just plan directly with this information? The sales forecast at each store is zero until the start selling date and the minimum stock requirement as defined by the planogram is effective a few days prior in order to allow store setup time. Not every store will necessarily have the same start selling date or lead time to its servicing DC, so allowing the product to flow as required to meet the merchandising need can actually have a smoothing effect on the extended supply chain.
What to do about items that are being allotted more or less space? Isn’t that just a function of the store’s future inventory position at the time the new planogram becomes effective? The future minimum display quantity is known, but the exact future on hand balance is not. Plug in what you know and continuously replan (and share) the needs from the supplier to the shelf until it’s time to commit.

Approving Orders

The universal truth is that committing to purchase product from a supplier is a risk. You are relinquishing the company’s cash to buy assets that may or may not provide a suitable return on investment between the time the product ships from the supplier until it ends up in the hands of the customer. It’s important to make sure that the decision is sound.

The universal assumption to address the risk is that a human being needs to approve purchase orders before they are released.

But…

Is the order quantity not the ultimate culmination of current decisions and assumptions around assortment, sales, pricing and stocking policies? If you’re comfortable with all of the inputs, then why would you be questioning the output? If you’re NOT comfortable with the inputs, then isn’t that where time should be spent?
If you’re planning on a continuum, then sales are always happening, inventory balances are always changing and the net requirements are always being adjusted. How does it help anyone to make 11th hour changes to plans that have been continuously evolving for weeks or months?

To be fair, these “sacred cows” have sprung up over time out of necessity. Retailers have been burdened with siloed planning and replenishment processes that provide no upstream visibility beyond the next order and – quite often – no forecasting capabilities at all. The vast majority are still burdened with them today.

That said, it’s 2018, not 1988 when a lot of those universal assumptions were made. Perhaps it’s time to start making some “sacred cow meatloaf” – I hear it’s delicious.

Franchisees and the Question of Control

Posted on April 4, 2018 by Jeff Harrop

I have a very strict gun control policy: if there’s a gun around, I want to be in control of it. – Clint Eastwood

A number of years ago, utilities companies, telecommunications companies, insurance companies, etc. started offering customers automated payment options as a convenience.

The premise was simple: Why bother taking the time to pay your bills each month when we can simply (with your permission, of course) automatically take the funds from your bank account or credit card?

When this was first offered, I must admit that I resisted for quite some time. I preferred receiving the bills each month, logging into my online banking site, selecting the payee and filling in the dates and amounts. I had control.

Month after month, my bills would arrive, each time with an insert or a perforated flap on the return envelope encouraging me to just give them my bank account or credit card number (and relinquish my control) to make my life easier. I continued to resist.

On occasion, I would make a dumb mistake entering the amount or the payment date. As a result, I would get dinged with late payment fees or other charges. I considered these penalties a small price to pay for retaining my control.

After years of managing my bills this way – and for no reason in particular – one month I came to the realization that not once had I ever decided to pay any amount other than the amount due on any other date than the payment due date. Not once. Ever.

It turns out that I wasn’t truly “controlling my destiny”. I was just wasting my time.

So I started signing up for automated bill payment with everyone who offered it. The bill would still arrive a couple weeks before the due date so that I could review the amounts and challenge any mistakes (i.e. I still had control), but no longer did I have to bother with all of the administrative nonsense to actually make payments.

With good reason, more and more retailers are moving in the direction of centralized store replenishment. Each day, POS data is collected and shared with the home office. And in many cases, the stores keep perpetual on hand balances and active planograms for every item. By planning store replenishment centrally, retailers can reap huge rewards by extending the planning process from stores to DCs to suppliers in a tightly integrated fashion (how does 98% in-stock while simultaneously dropping inventory sound?).

It’s really a shame that retailers with franchisee, owner/operator or dealership models can’t likewise benefit from centralized store replenishment. And why not, you ask?

Because the franchisees would never relinquish their control.

Hmmm…

Do the franchisees all use their own separate systems for store operations and staff their own I.T. departments? Does each store owner have his/her own fleet of trucks to pick up shipments? Do they each deal directly and independently with suppliers to bring product into their own dedicated warehouses? Does each franchisee develop his/her own advertising and promotion campaigns to draw customers into the stores?

You see what I’m getting at. There is a long list of value added services that a franchisor already provides to its franchisee store owners. Why couldn’t store replenishment also be such a service?

Like with automated bill payments, there are a few simple elements required in a franchisee model to ensure that independent store operators retain their control, while only relinquishing administrative tasks that add no direct value to customers.

Common Goals

From a customer standpoint, there is no line of delineation between the franchisor and the franchisee. There is a shared brand that receives the blame for poor customer service (or the accolades for a job well done). The needs of the end consumer must always be recognized as the raison d’etre for both the franchisor and the franchisee and be built into the supply chain processes like at any other retailer.

Transparency

Consumer centric rules of engagement need to agreed upon and executed faithfully at all times. For example, when there is a supply shortage at the DC, available stock needs to be rationed. Every franchisee wants to get his/her shipment, but when there isn’t enough inventory to go around, stores at risk of losing sales (and alienating customers) must have higher priority than those whose displays contain sufficient stock, but just won’t look very nice for a couple of weeks.

Where rules cannot easily be made, open and direct communication is a great substitute.

Visibility

Ronald Reagan once famously quipped: “The nine most terrifying words in the English language are: I’m from the government and I’m here to help.” Many independent store operators will have a similar reaction if the franchisor says “Trust us to handle your store inventory.” Franchisees are business people. As such, they will likely not be in favour of blind trust as a way to manage the businesses that feed their families.

Instead, franchisees will be better served if they can see the replenishment plans that are being managed on their behalf (much like getting your bills a couple weeks before the due date) and have the ability set their own policies to suit their individual businesses.

From the standpoint of meeting the most basic needs of the customer, there is no difference between a franchisee retail model and a corporate store retail model. Customers want stock availability and a good experience (however they define that) at a price they’re willing to pay. The fact that the ownership of inventory may have changed hands at the back door of the store before getting to the shelf is – to the customer – administrative and immaterial.

Why, then, should the process for managing the supply chain from the supplier to the shelf be any different under a franchisee model than for a retailer with corporate stores?

Software Selection: Who’s In Charge?

Posted on February 7, 2018 by Jeff Harrop

Decide what you want, decide what you’re willing to exchange for it. Establish your priorities and go to work. – H. L. Hunt

You’re living in a cramped 2 bedroom apartment and you decide it’s time to buy a house.

You have a spouse, 3 kids, 2 cars, a dog and a home-based business. On that basis, you determine that you’ll need a 2 storey home with 4 bedrooms, an office, a 2 stall garage and a medium-sized yard.

You give this information to your realtor and she compiles a list of houses for you to look at. Most of them are 2 bedroom bungalows with no garage on postage stamp lots.

In spite of your stated requirements, the realtor has determined that a 2 bedroom house is less costly to heat and cool, the cars can be parked outside and a small yard requires less effort to maintain. Most importantly, there’s an abundance of 2 bedroom houses on the market right now, so she’ll be able to get you into a house more quickly this way.

All of those things may be true, but you’re the one who’s going to be living there for the next several years, not the realtor. Sure, some tradeoffs are possible (even expected) – maybe the two youngest rugrats can bunk together until the oldest goes to college, or perhaps you can build your home office in an unfinished basement after you move in – but in this scenario, it seems as though the realtor is more interested in meeting her requirements.

All too often, a similar story plays out when companies embark on the journey of selecting new software to run their enterprise. We’ve seen it many times firsthand when it comes to supply chain planning software, but it really applies to any area of an organization that uses technology.

The burning platform may come from the business (“Our performance and productivity is suffering because of this clunky old legacy system!”) or from I.T. (“We can’t support this piece of crap any longer!”)

Regardless of the impetus for change, it’s the next decision where things start to go of the rails: “Well, since we need new software, then I.T. should be in charge of selecting it, right?”

Um, wrong.

I.T. certainly has a critical role to play, but if you’re the manager of a business process (and people) that will need to be supported by new technology, you take a back seat in the selection process at your own peril.

There are 3 things to always remember:

1. I.T. can’t read minds

I.T. folks love business requirements. Lots of them. With as much detail as you can muster. It helps them to shortlist candidates and research alternatives. If you don’t invest the time and energy to carefully think about what you want the future state to look like and write down your requirements, then you’re forcing I.T. to guess them. This would be like telling a real estate agent to ‘find me a house’ without providing any other details.

2. What I.T. considers important in a software package is not what the business considers important

Yes, the business has requirements, but so does I.T. A new software package must perform a number of functions in order to enable a new process. But it must also be supported (and supportable), process appropriate volumes and fit in with the established architecture of the systems that won’t be changing.

3. The business people will be using the new software each and every day to do their jobs

This speaks to the relative weighting between business requirements and technical requirements. The cheapest solution with the best processing speed and architecture that’s impossible for a user to understand and doesn’t support the business requirements isn’t much use. In other words, business requirements are more important and consequential to the running of the business than technical requirements – there, I said it.

So, suppose you get a good set of business requirements and a good set of technical requirements. A few software vendors come in to demonstrate functionality with the business folks and talk tech with the I.T. folks. You short list it down to 2 vendors:

Vendor A meets 80% of the critical business requirements, but 65% of the technical requirements
Vendor B meets 70% of the business requirements, but 75% of the technical requirements

What do you do now? The same thing you would do if the realtor at the beginning shows you a few houses that lacking in some way when compared to your ‘wish list’. You talk it through. You figure out which tradeoffs you’re willing to make. You discuss where workarounds might be used to bridge a functionality gap.

None of that is possible if you don’t give yourself the choice.