Logistics And Supply Chain Management

Where to Actually Start When You're Building a Supply Chain

Most people try to solve their logistics problems by buying expensive software they don't understand. That path costs more money and usually ends badly. I spent three years watching companies do exactly that before I learned how to approach this differently. The actual process is messier than any textbook makes it sound. Logistics And Supply Chain Management isn't one thing. It's a series of decisions about moving physical items from point A to point B, minus the cost of everything that goes wrong along the way. That last part — the "minus" — is where most operations bleed money. I still have spreadsheets from 2018 showing how much a single misrouted container cost one of my clients. The number was ugly enough that they fired their freight forwarder and hired someone off a referral.

Mapping Your Current State (Before You Change Anything)

Start by documenting what actually happens, not what your org chart says should happen. I had a client who claimed their average lead time was 12 days. The reality, once I pulled warehouse floor logs and compared them to carrier tracking data, was 23 days. The 12-day number came from a report generated by an ERP system that only counted days after the order left the dock. It completely ignored the 11 days orders sat in a staging area waiting for a container to be booked. You need the same honesty. Walk through every handoff. Document the actual time between each step. The numbers will surprise you. Then build your plan from those real numbers, not from the optimistic ones your sales team quotes to customers. The practical steps here are straightforward but tedious. Export your last 90 days of shipment data. Filter out any returns or damaged goods — they'll skew the averages. Calculate the median, not the mean, because one emergency air freight can distort the average significantly. Compare that against your quoted lead times. If the gap is more than 15 percent, you have a communication problem, not a logistics problem. Fix the communication first.

Choosing Your Freight Mix

There's a common assumption that you should always use the cheapest shipping option and then pay for expedited freight when things go wrong. That strategy sounds reasonable until you calculate the total cost of expediting, which usually runs three to five times the base rate. I saw a mid-market distributor lose $40,000 in a single quarter doing exactly that. Their "savings" from ground freight got eaten by ten separate expedited shipments. The better approach is to segment your SKUs by velocity and margin. Fast movers with healthy margins can absorb premium freight without hurting profitability. Slow movers with thin margins need a different strategy entirely — you're better off holding slightly more inventory locally than paying for expedited ground or air on a product that makes $2 per unit. The math doesn't work either way. Carrier selection matters less than people think, as long as they're reputable. The difference between a solid carrier and an excellent one on a given lane is usually less than 8 percent in transit time variation. Spend that energy on lane optimization instead. Consolidating two LTL shipments into one FT can save you 30 to 40 percent on the line-haul portion. That's a real number, not an estimate.

The Warehouse Problem Nobody Talks About

People focus on transportation because it's visible. What actually kills most operations is warehouse flow. I worked with a company whose shipping errors were running at 4.2 percent. The root cause wasn't their WMS or their pickers. It was that they had put fast-moving SKUs in the back of the warehouse to free up front space for a seasonal display. Pickers were walking longer distances, getting tired, and making mistakes. Moving those SKUs to the front cut errors to 1.1 percent in two weeks. No technology change. Just reorganization. When you design or redesign your warehouse layout, measure the average travel time from staging to dock during a typical shift. If it's over 15 minutes per order, you have a layout problem. Cross-docking can eliminate that entirely for high-volume items, but it only works if your receiving and shipping schedules are synchronized. If your inbound trucks arrive at random times, cross-docking becomes a disaster because you can't guarantee the outbound truck will wait or that the product will be staged in time. For smaller operations, just zone your warehouse by velocity. ABC analysis done properly — using trailing 90-day movement, not annual — will show you which items belong where. The top 20 percent of SKUs typically account for 80 percent of the picks. Put those within 10 feet of the packing station. Everything else can live further back.

Inventory Planning Without the Certainty You Don't Have

Beginners try to predict demand. That's not how this works. The people who survive do something else: they build buffer into the places where predictions are least reliable and let the predictable parts run lean. Safety stock isn't a punishment for poor forecasting. It's insurance against the fact that your forecast will be wrong, always. The standard formula for safety stock is Z times the square root of lead time times the standard deviation of demand. That formula assumes demand and lead time are normally distributed and independent. They're not. I've seen companies use this formula with data that was heavily skewed by promotions, stockouts, or seasonal events, and then wonder why their safety stock was either way too high or dangerously low. The fix is to normalize your data first. Remove promotional spikes. Backfill any stockout periods with estimated demand. Then run the calculation on clean data. Another common mistake is setting reorder points based on a single supplier's lead time. If you have two suppliers, calculate separate reorder points for each. The one with the longer or more variable lead time needs a higher buffer. I had a client who kept running out of a component because their secondary supplier had a 22-day lead time variance of plus or minus 5 days, but they were ordering using the primary supplier's 10-day lead time. They were perpetually one bad week away from a stockout.

Technology Choices That Don't Suck

Most ERP implementations fail because they're designed for how the company wants to work, not how it actually works. The ones that succeed start with a clear definition of what the system needs to do, written in operational terms. "Track inventory" means nothing. "Show me, in real time, how many units of SKU 4471 are available at the Chicago DC, in transit from the supplier, and committed to open orders" means something. Write your requirements like that. If your operation is small enough that a full ERP is overkill, a decent WMS paired with a basic TMS and a shared spreadsheet for demand planning will handle most mid-market volumes. The integration between those tools should be manual at first. Don't automate bad processes. Map the workflow on paper, prove it works for a month, then build the automation. I've seen companies automate a broken process and then spend more time troubleshooting the automation than they would have just doing the process manually. For transportation management specifically, start with a tool that handles rate shopping and lane consolidation. Don't bother with predictive ETA engines until you have at least 500 shipments per month. Below that threshold, the data isn't dense enough for those systems to be accurate, and you'll waste money on a feature you can't validate.

When Things Break (And They Will)

Last year, a port strike in Southeast Asia disrupted a client's primary supply lane. Their backup supplier was in a different country and couldn't fill the gap for six weeks. They'd never modeled that scenario because their risk assessment only looked at single-point failures, not correlated failures across their entire network. The fix was simple in hindsight — identify all your critical components, map every supplier and their geographic region, and flag anything that shares a port or corridor with another critical component. Then build a contingency plan for each flagged combination. That took them about four hours. The alternative was losing two weeks of revenue while they scrambled. I'd rather spend four hours on that exercise than deal with the aftermath again. Another pattern I see repeatedly: companies optimize for cost per unit shipped and forget about cost per unit sold. A freight rate that looks cheap on the surface might come with minimum order quantities that force you to hold excess inventory. The carrying cost of that inventory — warehouse space, capital tie-up, obsolescence risk — often exceeds the freight savings. Do the full calculation. I once found a "cheap" ocean rate that was actually 18 percent more expensive on a landed-cost basis once I accounted for the extra inventory carrying cost.

What This Approach Won't Do For You

None of this eliminates variability. Supply chains are inherently unpredictable. Weather, politics, supplier bankruptcies, and demand spikes will keep happening. The goal isn't to prevent disruptions. It's to reduce their impact and recover faster than your competitors. This also requires honest data. If your company treats inventory counts as suggestions rather than facts, none of these methods will produce reliable results. Cycle counting isn't optional. I've seen companies skip it for years, then wonder why their demand planning was producing garbage outputs. The garbage-in-garbage-out principle applies here with brutal consistency. Finally, this work is ongoing. The moment you stop reviewing your metrics — even monthly — your operation starts drifting. A lot of companies set up their logistics framework and then forget about it until something breaks. That's when you learn what your blind spots are, usually at the worst possible time.