Grand experiment on the TCO of a supply chain software
The price tag1 of a piece of software ranges from nothing, as it happens with open source, to quite a lot - enterprise software leaning heavily towards the latter. However, software doesn’t operate in a vacuum, and operating a piece of software always involves some degree of overhead. The notion of TCO (Total Cost of Ownership) precisely tries to capture the cost in full, taking into account both direct and indirect costs. From a corporate perspective, it’s the TCO that matters, not the price tag. Naturally, the TCO cannot ever be lower than the price tag.
The price tag of Lokad2 frequently comes as a surprise to our prospects. However, I am defending in this post that Lokad is one of the lowest-priced options of the market, irrespective of the return on investment. Naturally, this exercise is fraught with biases, but I invite my peers to respond in kind as there is no such thing as neutral observers in the realm of enterprise software.
Supply chain-wise, the one cost factor that dominates all others is the headcount required to keep orchestrating the flow3. The list of related job titles is long: inventory manager, supply planner, demand planner, production planner, category manager, purchase manager, forecast manager, planning manager, supply chain analyst, supply chain professional, S&OP manager, to name a few. Job titles vary from one company to the next, but all those positions have something in common: They are all dedicated to the ongoing orchestration of the flow of physical goods, but not physically operating the flow. If we had Strong AIs (we don’t, not even close), this headcount would be zero.
However, the Lokad approach, when taken to the limit, should, in theory, achieve a near-zero headcount on the company side. We don’t have AIs, but our subscription comes with a team of supply chain scientists who take it upon themselves to deliver the end game4 flow orchestration. We pride ourselves in making sure that our numbers are final, and that there is no need to go through another round of Excel tweaks before letting those suggestions come to pass as orders (purchase orders, production orders, stock move orders, repair orders, etc).
Until 2020, this near-zero headcount remained mostly aspirational. Our clients were maintaining a buffer of supply chain employees, if only to continuously audit and challenge the numerical recipes designed and maintained by the Lokad teams. However, the pandemic happened, and things started to get weird.
In 2020, several countries across Europe started to enforce lockdowns while offering near full pay compensation for employees who were forced to stop working - as if they were on sick leave. Many companies, assessing that times ahead would be rough, took advantage of those measures to the fullest extent. They put almost their entire white-collar workforce “on leave” while keeping employees on the ground - deemed “essential workforce” - operating more or less as usual.
For quite a few of our clients, the supply chain scientists at Lokad were suddenly on their own, while supply chains were still effectively in operation5. This situation lasted about 14 months. During this period, Lokad was de facto in charge of the daily orchestration of over 1 billion euros worth of inventory, with little recourse as client teams were on leave.
Overall, it went surprisingly smoothly, especially considering that countries across Europe were opening and closing erratically, changing their strategy every couple of weeks, giving companies less than a week of leeway to adjust to a new set of far-reaching policies6. Our team of supply chain scientists spent most of 2020 and 2021 keeping our numerical recipes under constant revision, just to cope with the ever changing market landscape.
The events of 2020 and 2021 proved that the TCO of Lokad was essentially equal to our subscription cost. The proof came from an accidental response to policies outside of our control. Neither our clients nor Lokad had any plans to carry out such “grand experiments”, but we did, and they went fine7. I do not regard those events as proof that it’s reasonable to leave the supply chain operated entirely by Lokad. Keeping a team around to challenge Lokad is wiser. In short, trust, but verify.
In contrast to this perspective, let’s have a look at the other supply chain software vendors. The vast majority of my competitors are charging per user8 or something roughly aligned with headcount. The fundamental issue with this approach is that the vendor is directly incentivized to make the solution as unproductive as possible. The more employees the client company has to throw at the problem, the more the vendor benefits.
Unsurprisingly, competing solutions offer low productivity: typically, between 1 and 2 orders of magnitude lower than Lokad when looking at a metric such as SKUs per planner. The very existence of capabilities like Alerts and Exceptions - which are absolute time sinks, as they mandate manual interventions from the supply chain teams - should be a big red flag, but somehow many of my competitors are touting those features as if they were “a good thing”9.
As far as the TCO is concerned, every single employee that interacts with the piece of software should be accounted for. For jobs like “planners” or “inventory managers”, as a rule of thumb, it’s reasonable to assume that 100% of the full employer cost belongs within the TCO envelope of the software. With larger companies, the TCO can get even bigger as “forecasting” or “data science” teams get introduced to supplement the efforts of the planners. Those extra teams should also be put within the TCO envelope10.
Cost-wise, those jobs invariably dwarf the monthly subscription costs of Lokad tenfold - and frequently more. Surprisingly, this is usually the reason why, despite the introduction of Lokad, companies do not proceed with headcount reduction: overall, Lokad barely moves the needle in the grand scheme of the complete supply chain budget.
However, it’s not because the company hasn’t yet decided to shrink its supply chain headcount, that the supply chain vendor can walk away scot-free. There is an army of clerks that happen to be effectively mobilized by the piece of software that has been deployed. The vendor can only be credited with a low TCO if the team can effectively turn its attention elsewhere, at the very least for a couple of months.
Thus, TCO-wise, I have yet to see any supply chain vendor that does happen to be many times more expensive than Lokad. If a low price tag is what you are looking for, then I can only recommend Microsoft Excel which comes at a price of just a few dollars per month for the user.
I am referring to the amount charged upfront by the vendor to the client company. This can take the form of licence fees, maintenance fees, or any kind of subscription fees. ↩︎
A monthly subscription with Lokad typically involves a platform fee plus a service fee. The platform fee is basically an all-inclusive package for a programmatic environment delivered by PaaS (Platform as a Service). The service fee covers the dedication and experience of a supply chain scientist from Lokad. ↩︎
Supply chain orchestrates the flow of physical goods, but it does not execute the flow. It should not be confused with logistics (physically moving the goods), production (physically transforming the goods) or distribution (physically servicing the goods). ↩︎
For every single SKU, there are a series of questions to answer, such as: How much do we buy? How much do we produce? How much for the selling price? Do we move the stock? Do we scrap the stock? etc. ↩︎
Aviation MROs were among the hardest hit with -30% to -50% in volume. However, for ecommerce, volumes were frequently at +20% to +50%. Management was frequently planning for the worst. We had several ecommerce clients who put their white-collar workforce on paid leave - the government footing the bill - while volume was surging. ↩︎
Policies were quite innovative as well. For example, in several countries, at some point, stores were allowed to open, as long as the sale was performed outdoors. Thus, Lokad had to devise inventory policies for stores that would only temporarily operate with a reduced assortment exposed on the sidewalk in front of the actual store. ↩︎
I would be tempted to say it went “great”, but according to what? The situation was fully unprecedented. We still don’t have any baseline for any meaningful comparison. During the odd period where Lokad was left to its own devices, I only received a handful of calls from alarmed supply chain directors, and in the end, we did not generate any notable incidents while opportunities for incidents were pouring in. ↩︎
Lokad does not charge per user. In our service agreements, we do put limits on the number of users - typically twice higher than current headcount - but only as a safeguard. The intent is to clarify the scope of our commitment, not to engineer an upsell mechanism. The limit is expected to be high enough not to be reached, taking ≈5 years of company growth into account. ↩︎
In enterprise software, the pinnacle of marketing mastery appears to be, being able to present the worst flaws of a product under the guise of desirable qualities. ↩︎
If the company needs to maintain a team of data scientists to supplement the defective numerical recipes offered by the vendor in charge of supply chain optimization, the vendor is clearly responsible for this situation. ↩︎