Innovative Solutions for Distribution: Lanthorn in Construction

In today's fast-paced world, efficient distribution and manufacturing processes are crucial for the success of any business. Below is a real-life case study that we completed for a construction materials distributor.

The Basics: Our customer had 4 distribution centres across the country that was serving demand for 278 different products. The following practices were in place:

• Order Timing & Quantity: Orders are placed to meet only the current period’s demand, with no future forecasting or rolling-horizon planning.

• No Shortage Penalty Consideration: Stockouts are tolerated, with no penalty or lost-sales cost considered for unfulfilled demand. Shipments are not expedited, leading to frequent missed sales.

• Single Slow Transport Mode: All orders use the cheapest, slowest transportation mode, ignoring faster options that could prevent stockouts.

• No Advance Inventory Buildup: Inventory is replenished reactively after demand occurs, with no strategic buildup or safety stocks for future periods. Extra inventory only results from operational constraints like MOQs.

• Minimum Order Quantities (MOQs): MOQs apply to orders. If demand is below the MOQ, the planner can delay ordering (risking stockouts) or order the minimum and carry extra stock. This often results in stockouts for low-demand items or small surpluses, without a safety stock strategy.

• No Demand Forecasting: Demand forecasts or trend analysis are not utilized. Initial inventory at DCs is based on current stock, possibly from previous guesswork. Decisions each period do not account for future demand changes, leaving the system unprepared for sudden demand spikes.

• Ordering Pattern: The planner orders only the exact demand for each period, adhering to MOQs, without consolidating orders across periods. This approach can lead to under-utilized transport capacity during low-demand periods and last-minute adjustments in high-demand periods.

  
  

This approach reflects classical “fire-fighting” approach, which often minimises cost in the short-term but at the expense of service level and long-term efficiency. This provided an opportunity for Lanthorn to showcase what optimisation could offer in both serving costs while increases servicing of key customers, enabling higher revenues as a result.

Overall Performance Summary:

Below is a summary of key metrics over the 12-period, 4-DC, 278-product case:

Metric (12 periods total)Spread	Baseline	Optimised
Total Demand	1,667,000	1,667,000
Units Fulfilled	~1,170,000	~1,335,000
Units Short (Unfulfilled Demand)	~497,000	~332,000
Service Level (Fill Rate)	70%	80%
Transportation Cost (est)	~$1.0M	~$2.0M
Shortage Penalty Cost (lost sales)	~$50.0M	~$33.3M
Total Cost (Transport Shortage)	~$51.0M	~$35.5M
Peak Inventory (units on hand)	~98,000 (late, by period 12)	~228,000 at Period 6
Average Inventory	97,500	111,250
Shortage Concentration	Huge early spikes and recovery	Spread ~20% over each period.

Below are is the summary of the key decisions we helped enable through optimisation:

Total Cost Breakdown: Transportation vs Shortage Penalties

The original scenario of distribution exclusively used cheap shipping options, which enabled lower cost of transport-but often at the expense of lost sales due to not arriving within customer requirements. These lost sales, while currently quantifiable in lost revenue, represent a larger land-mine of reducing customer trust, relationships, and even opportunities for competitors to enter the market.

What did we do? Well we quantified this lost sale in our model, which generated a significant penalty for the model if we do not fulfil demand-often pegged at a much higher cost than transportation. This resulted in a total cost of about $35.3m of which 95% was penalty of unfulfilled demand, with transportation costs only coming in around $1.7-$2m for the total 12 months forecasted. This represents a model that wisely prioritised fulfilling orders within constraints, as each unit shipped saved a $100 penalty-which would have equated over $133m in penalties. Unfortunately, even the model could not fulfill all demand- with 332k units not being able to be filled, resulting in $33.25m penalty. Results are graphed below:

Why quantify lost sales? Well, in the original scenario (without a lost sale penalty) we expected a transport cost of ~$1m, but a potential order loss of $50m.

In summary, the optimised model achieves a far better cost-service balance: it deliberately spends more on shipping (using faster modes when needed) to dramatically reduce the larger cost of shortages. The baseline approach, lacking this strategic trade-off, pays the price in unmet demand. This stark cost breakdown underscores the value of advanced planning – every dollar in freight can save ~$20–$30 in penalty costs in this case, a point that the baseline practice misses.

Inventory Profile over time:

Another key difference is how each approach manages inventory over the 12 month optimisation period.

Below is a graph of Inventory on hand (total across all DCs) at each period for optimised (yellow-line) vs original (orange-line) scenarios.

The optimised scenario intentionally builds up inventory in early periods and then draws it down over time. Orginal scenario starts with initial stock available and then runs it down to zero by period 2 since no replenishment arrived in time- after which inventory build is overbuild to compensate. This relfects bull-whip effect: initial under-stock leading to shortages, followed by catch-up ordering and inventory accumulation later.

In the optimised model, inventory was used as a strategic tool. The optimiser anticipated a mid-horizon demand spike and pre-stocked the network in advance thanks to forecasting and the rolling horizon. By the end of period 1, it kept a modest 22,700 units in stock across the DCs. It then increased stock levels each period, reaching a peak of ~228,000 units by period 6. Each of the 4 DCs saw a similar peak of ~55–58k units then. This was intentional: the model-built buffers early so that when demand spiked, inventory was on hand to meet it despite long transport lead times and MOQS. Essentially, the optimiser “filled the pipeline” early, then relied on that stock to satisfy later orders, ending with moderate leftover. Notably, no product was massively overstocked – the most extensive ending inventory of any single item was ~1,543 units, and most had only small leftovers. This balanced profile (build-up then taper down) demonstrates proactive inventory smoothing: the optimised plan prevents shortages by holding inventory in advance and avoids bloating stock unnecessarily. We can best see this when we look at its outcomes- inventory shortages:

From an operations perspective, the optimised inventory strategy clearly outperforms. Smoothing inventory over time ensures product availability during peak periods (no DC ever hit zero stock before the spike). The original scenario’s lack of safety stock meant it could not buffer any disruption or surge – inventory hit zero in period 2, resulting in missed sales. Although the baseline eventually rebuilt inventory, it did so too late; those late arrivals became excess stock rather than serving the customers who needed the product earlier. In short, the optimised model uses inventory proactively as an asset, while the original treats inventory purely reactively, leading to instability. As research shows, a forecasting, rolling-horizon approach prevents extreme inventory swings and stockouts, which is exactly what we observe.

Conclusion: Operational and Financial Impact

The comparison between the optimised supply chain model and the original baseline scenario highlights significant operational and financial impacts:

• Service Level & Reliability: The optimised model provided a significantly higher fill rate (~80% vs ~70%) with a far more consistent performance across periods, products, and regions. This translates to more dependable customer service and higher revenue capture. The original approach suffered unacceptable stockouts (up to 80% of demand lost in a period) and highly uneven service, undermining customer trust.

  
  

• Cost Efficiency: Counterintuitively, the optimised scenario, despite spending more on transport (expedited shipping), achieved a lower total cost by drastically reducing the enormous cost of shortages. The baseline’s laser focus on minimising transport cost backfired – it led to such extensive lost sales that total costs (when accounting for those penalties or lost profits) were much higher. In short, the optimised supply chain was much more cost-effective when considering the big picture; it invested in prevention, whereas the original paid for failure.

  
  

• Inventory Utilisation: Operationally, the optimised plan used inventory as a strategic buffer – increasing stock when needed and drawing it down efficiently. This prevented disruptions and met demand peaks. The original plan’s reactive inventory management resulted in the classic bullwhip pattern: it saved inventory holding early on but at the cost of stockouts and then held excess inventory later anyway. The optimised network had inventory in the right place at the right time, while the original often had inventory at the wrong time (either too late or not at all when needed).

  
  

• Order and Supply Stability: The optimised approach smoothed out orders and fully used available transport capacity, leading to a more stable supply chain. Production or procurement upstream could be more predictable because the orders were placed with foresight. The original approach likely led to erratic order patterns (periods of no orders followed by rush orders), which can strain suppliers and increase unit costs. By batching orders logically and avoiding last-minute expedites (except when necessary), the optimised model improves supplier relationships and lowers per-unit costs long-term. The baseline’s ad-hoc ordering can cause expediting by crisis (even if not via faster mode, a late large order can force overtime or emergency production at a supplier) and inefficient workflows.

  
  

• Prioritisation and Market Strategy: The optimised plan inherently aligns with business strategy – it ensured high-value products and all regions were well-served. This means the company can meet its strategic sales objectives (e.g., not missing out on premium product sales, keeping all areas supplied). The original scenario had no such alignment, risking strategic failures (like running out of a flagship product). This could damage the brand and allow competitors an opening. Financially, the optimised scenario’s ability to fulfil more high-margin demand likely improved profitability mix, whereas the original might disproportionally lose high-margin sales and only make low-margin ones (since those might be the only ones that didn’t stock out).

  
  

• Risk Management: The optimised approach is far more robust to uncertainties. Even when faced with the same demand variability and constraints, it handled them in a controlled manner. The original approach had no cushion – any unexpected surge immediately caused a failure (stockout). In supply chain terms, the optimised model had resilience built in (via safety stocks, flexible shipping), whereas the original was brittle. The financial impact of that is seen in the cost of shortages and potential long-term loss of market share if customers switch due to unreliability.

In conclusion, the optimised supply chain model dramatically outperforms the original “naïve” scenario in operational metrics and financial outcomes. Smartly balancing inventory and transportation achieves a higher service level at a lower effective cost. The Original scenario, reflective of poor planning practices, may appear to save money on paper (cheaper shipping, less inventory). Still, it incurs much larger hidden costs in the form of lost sales, emergency fixes, and customer dissatisfaction. The case study demonstrates that investing in optimisation – forecasting, rolling horizon planning, and modal choices – yields tangible benefits:

A)      Roughly 10% more demand fulfilled,

B)      ~$15M cost avoidance in our 12-period example,

C)     and a smoother, more reliable supply chain operation overall.

These improvements align with industry best practices and benchmarks, which show that companies employing advanced planning achieve lower total costs and higher fill rates. Stakeholders should take away that moving from the “original” reactive approach to an optimised approach is not just a theoretical exercise, but one that drives significant competitive advantage and value for the business.

If you want to learn more about how we can optimise distribution for your business, reach out to hello@lanthornsc.com !