Category: ERP

Twelve-day closes are usually not a staffing problem. A $45 million manufacturer I worked with had five people in finance, and three of them lost the first weekend of every month to the close. When we asked what would happen if the close took four days, the controller laughed. Then she started listing reasons it couldn’t.

We wrote the reasons on a whiteboard. There were seven. Six of them were self-inflicted.

The four things eating the calendar

We tracked where the hours actually went during one full close, and the picture was lopsided.

• Waiting on inventory. The warehouse counted every month-end and delivered numbers on day three, so finance sat on its hands for 60 hours.
• Freight invoices arriving after cutoff. About 300 invoices a month landed with dates from the previous period, and each one meant a manual adjustment.
• Nine spreadsheets reconciled by hand. Each took two to three hours, and at least one had a broken formula nobody had noticed.
• Approvals sitting in inboxes. Expense and journal approvals averaged 2.4 days to come back.

Only one item on that list was a real constraint — the physical count. The other three were habits.

Start with the last three days, not the first

Most teams try to speed up the beginning of the close. That’s the wrong end. The last three days are where the rework lives, and rework is where the time goes.

1. Move the physical count off month-end. Count the highest-value items weekly all month so the month-end figure is a confirmation, not a discovery.
2. Set a hard cutoff for freight and receiving documents at 5 pm on the last business day. Anything later goes to the next period and gets disclosed, not adjusted.
3. Replace one spreadsheet at a time. Pick the worst one, rebuild it inside the system, and run both for one close before you retire the old file.
4. Give approvers a fixed 15-minute window every morning. Everything above $5,000 clears in one sitting.

This is the kind of work where ERP Consulting pays for itself in weeks rather than years, because someone has to sit with the controller and argue about the cutoff rule until it sticks.

What the team got back

By the fourth close they were done in four and a half days. By the sixth, four. Nobody was hired, and nobody worked a weekend.

The bigger win wasn’t the calendar. It was that the numbers arrived while they still mattered. Sales could see month-to-date margin on day five instead of day thirteen, and the owner stopped making pricing decisions on nine-day-old data.

One useful ERP Consulting rule of thumb: never automate something you haven’t simplified first. Every hour spent deleting a step is worth three hours of configuration.

If your close runs past eight days, do this before you change anything else. For one month, have everyone log what they were waiting on, in 30-minute blocks. The answer is almost always waiting, not working — and waiting is fixable.

The sales manager walked into the Thursday morning meeting and said four words: “We lost Riverside.” Riverside was a $400,000 fabrication job, the kind that keeps a shop busy for three months. The customer’s reason was one line long — “we asked for pricing 11 days ago and never heard back.”

Nobody in that building was lazy. The estimator had done his part in a day and a half. The problem was everything that happened between the parts.

A quote that needed four people and three waits

This was an 85-person metal fabricator doing about $19 million a year. Any quote over $50,000 needed engineering to confirm tolerances, purchasing to price the steel, and the shop foreman to estimate hours. Each person finished their piece in a day or two. The waiting in between ate the rest of the week.

• Engineering finished Tuesday and emailed the file. Purchasing didn’t open it until Thursday.
• Steel pricing came back from a supplier in 36 hours and then sat in an inbox over the weekend.
• The foreman’s hours estimate was given verbally in a hallway and written down by nobody.

Add it up and the average quote took 8.6 days to leave the building. The win rate was 22%.

One rule fixed most of it: somebody owns the clock

They didn’t need faster people. They needed one person whose job was to know where every open quote was sitting and how long it had been sitting there. That’s the entire idea.

1. Name a quote owner, and it isn’t the salesperson. It’s a coordinator who sees every request from arrival to send.
2. Put a clock on each stage. Engineering gets 24 hours, steel pricing 24 hours, shop hours 24 hours. Miss it and the coordinator escalates that afternoon, not next week.
3. Capture the estimate where it’s used. The foreman’s hallway number goes into the quote the moment he says it, not into someone’s memory.
4. Send partial quotes when the customer is waiting. If steel pricing is the only holdout, send everything else with a clear note about what’s pending.

Notice what’s missing from that list: software. The tool matters far less than the clock. When ERP Consulting teams get this right, they spend most of their time on the rules and almost none on the screens.

What changed in 60 days

Average turnaround dropped from 8.6 days to about 31 hours. The win rate moved from 22% to 34%, partly because they were faster and mostly because they stopped being forgotten.

The revenue number is the one that got the owner’s attention: roughly $2.9 million in additional annual bookings from the same estimating team, the same shop capacity, and the same customer list.

A second ERP Consulting habit stuck as well. The coordinator now starts every morning by listing quotes that are more than two days old and calling whoever is holding them. It takes ten minutes and it has never once been a wasted call.

Try this before your next sales meeting. Take your ten most recent quotes that took more than five business days to send. Write down where each one sat the longest and who was holding it. Nine times out of ten, the same two handoffs show up. Fix those two first.

Ask a distributor how accurate their inventory is and you’ll usually hear “pretty good, maybe 95%.” Then you count. A 60-person parts distributor I worked with ran a full physical count last March and came up 12% short — on $6.2 million of stock, that’s roughly $740,000 of product that existed only on a screen. Nobody stole anything. It leaked out one small error at a time, over four years.

The owner’s first instinct was to blame the system, and his second was to buy a new one. Both were wrong, and realizing that saved him about $180,000 in software he didn’t need.

The three places inventory quietly disappears

We traced the 200 worst discrepancies by hand, and every one of them fell into three buckets. None of it was exotic.

• Receiving shortcuts. A truck pulls in at 4:40 pm, someone signs for 40 cartons, the paperwork said 42. Twice a month, across 60 suppliers, that adds up to thousands of units a year.
• Substitutes on the pick floor. A backordered part gets swapped for the closest match and the paperwork never follows. The customer is happy. The count is not.
• Returns parked in the corner. Nine pallets of customer returns were booked back into stock before anyone inspected them, and about a third were unusable.

What fixed it, and it wasn’t a new system

We brought ERP Consulting in for the counting routine and left the software alone. The fix started with counting more often, not less. Most teams count once a year because it hurts, which is exactly why it hurts.

1. Sort the catalog into three groups. The top 200 parts by value get counted weekly, roughly 40 items a day. The middle group monthly. The long tail twice a year.
2. Give one person per zone the credit and the blame. Shared accountability across an aisle turns into no accountability.
3. Count the shelf, not the screen. If a counter sees “should be 14,” they will find 14. Blind counts only.
4. Chase the cause the same day. Every miss over $200 gets a one-line note — wrong pick, bad receipt, damage. After six weeks the top three causes were obvious.

This is where ERP Consulting actually earns its fee, and it’s rarely the part people expect. Nobody needs help switching on a counting screen. They need someone standing next to the receiving crew at 4:40 pm asking why the number drifted.

Day 90, in numbers

Accuracy hit 94% by week six and 98.6% by week twelve. Shrinkage fell from $740,000 a year to under $90,000. The company also stopped paying for expedited freight on parts it thought it owned.

The habit that held it together came out of a follow-up session: every Monday, the warehouse lead and the controller looked at one page — the five worst items from the past week — and fixed the cause, not the count. Fifteen minutes, and that’s the whole meeting.

What surprised the CFO most was that customer service improved instead of slipping. Fewer “sorry, we don’t actually have that” calls, fewer credits, and a warehouse team that stopped getting yelled at every Friday.

Here’s something you can run this week. Pull the last 20 order lines that had to be shorted or substituted even though the screen showed stock. Hand those 20 item numbers to the warehouse and ask for a blind paper count. If more than two come back wrong, your next quarter has already written itself.

A specialty chemicals distributor with 220 employees had a rule that any purchase over $1,000 needed the owner’s approval. It made sense in 2009, when they had 30 employees and $4 million in revenue. Last year they had $52 million in revenue and 4,100 purchase orders waiting in one person’s inbox.

Average approval time was 3.6 days. The owner was the bottleneck and he knew it. He just didn’t know how to loosen the rule without losing control of spending.

The real cost of a slow signature

We pulled a year of purchase orders and sorted them by size. The picture was almost comical.

• 78% of POs were under $2,500 and totaled just 6% of spend.
• Roughly 1,900 of them were routine consumables — gloves, solvents, packaging film.
• Fourteen POs accounted for 41% of total spend.

He was spending his attention on the small stuff and had almost none left for the fourteen orders that actually moved the business. Approval rules that don’t scale stop being controls and start being delays.

There was a hidden cost too. Buyers, trying to dodge the wait, were splitting orders — two $900 orders instead of one $1,800 order. Double freight, double receiving, and a supplier who noticed and started pricing accordingly. The rules were built for a company one-tenth the size, which is the classic pattern an ERP Consulting review uncovers first.

How they rebuilt the approval ladder

1. Under $500: automatic, no approval, if the item is on the standard list. About 40% of transactions left the queue overnight.
2. $500 to $5,000: buyer’s manager approves, 24-hour clock, escalates on its own if missed.
3. $5,000 to $50,000: department head plus finance, against a budget already agreed.
4. Above $50,000 or any new supplier: owner signs. That’s about 40 decisions a year instead of 4,100.

They also stopped order splitting cold by flagging any two POs to the same supplier within seven days. Two buyers got a conversation. It stopped.

What the owner actually got back

Routine purchase approvals fell from 3.6 days to under six hours. Expedited freight charges dropped by $71,000 in the first year, mostly because buyers no longer had to pay rush shipping to cover internal delays.

The number he cared about was different, though. He went from reviewing 4,100 transactions to reviewing about 40, and those 40 carried the real money. He told me it was the first time in four years he felt like he was running the business instead of clearing an inbox.

This month, export your last 500 purchase orders and sort them by amount. Count how many were worth the executive attention they received. For most managers, the top 5% of orders drive the majority of spend — and the other 95% are just clogging the pipe.

Two equipment rental companies. About 180 machines each — skid steers, lifts, compact excavators. Same metro area, overlapping customers, nearly identical rates. Last year one ran a 19% operating margin. The other ran 13%. The difference traced back to three decisions.

The first difference: they knew which machines earned money

Company A tracked utilization by individual machine, by month — hours rented divided by hours available. Company B tracked it by category, at year end, from memory and gut feel.

When Company A looked at the real numbers, the finding stung. Their 12 boom lifts averaged 71% utilization with strong margins. Their eight 60-foot telescopic lifts averaged 22% utilization and quietly ate $180,000 a year in maintenance, storage, and financing. They sold four and stopped buying more.

Company B still had all of theirs and no idea, because nobody had ever asked the question. That question is the whole job in most ERP Consulting engagements I’ve watched. Equipment you’re not sure about is usually equipment you should sell.

The second difference: turnaround time between rentals

Every machine that comes back needs washing, inspection, and repair before it goes out again. Company A’s average turnaround was 1.4 days. Company B’s was 4.2 days.

That gap alone is worth about 8 percentage points of utilization on a machine rented 300 days a year. Company B’s team wasn’t lazy — they found out what needed fixing when the next customer called, instead of when the machine rolled back in.

A one-page checklist at return, logged in the same place the rental contract lives, closed most of the gap in about ten weeks. No new equipment, no new hires.

The third difference: how they priced the awkward jobs

• Deliveries beyond 25 miles got a separate line item instead of being absorbed into the daily rate.
• Weekend rentals priced by day, not by “a weekend” — a Friday-to-Monday rental is three days of wear.
• Damage deposits collected up front, with photos taken at delivery and at return.
• Repeat customers got a discount on rate, never on delivery or damage.

Those four changes added roughly $290,000 in annual revenue at Company A without a single increase to the core rental rate. Company B was giving all of it away as a favor and calling it customer service.

If you rent or lease anything — equipment, vehicles, instruments — find the worst 10% of your assets by utilization and decide this quarter whether they stay. You don’t need a new system. You need one report and the willingness to act on it.

At 6:40 on a Thursday morning, a food processor got the call every plant manager dreads. A customer found a foreign object in a finished case. The retailer wanted to know how many other cases came from the same batch, and they wanted the answer before the store opened.

This company answered in nine hours. Eighteen months earlier, the same scenario would have taken nine days and probably cost them the account. Here’s what changed in between.

The question you cannot answer slowly

Traceability is not a paperwork exercise. When a recall hits, three questions decide everything: which raw material lots went into this batch, what finished goods came from those lots, and where those goods are right now — plant, truck, warehouse, or shelf.

Their old answer lived in three places: a paper receiving log, a production binder, and a shipping spreadsheet. Reconstructing one batch took a person two full days. If your trace takes longer than a day, you’re choosing between a recall and a reputation.

What we captured, and what we deliberately skipped

The temptation is to track everything. That fails, because people stop recording when recording gets tedious — a pattern any honest ERP Consulting review will show you in the first week of interviews.

We narrowed it to four data points, captured at the moment the work happened.

• Supplier lot number, scanned at receiving — not typed later.
• Batch number created when production started, tied to the lots consumed.
• Case and pallet IDs printed at packaging, tied to the batch.
• Shipment record linking pallet IDs to customer and date.

That’s four links in the chain. The whole thing runs on a handheld scanner a line operator already knew how to use. Training took one shift, not one week.

The nine-hour Thursday, hour by hour

1. Hour 1: Customer service pulled the case code. The system returned a batch number and a production date.
2. Hours 2–3: The batch record listed two raw lot numbers and every pallet produced that shift — 1,840 cases.
3. Hours 4–6: The shipping record showed where all 1,840 cases went: 1,100 to one retailer, 620 still in the warehouse, 120 in transit.
4. Hour 7: The retailer got an exact case count and a list of store locations. No guessing, no over-recall.
5. Hours 8–9: The warehouse team quarantined the 620 cases and held the in-transit truck with one phone call.

The retailer was impressed enough to expand the account three months later. The recall cost about $14,000 in product. A nine-day version of the same event would have hit six figures and made the local news.

Next week, pick one finished product in your warehouse and trace it backward to its raw materials using only your records. Time yourself. Whatever slows you down is your first project.

Most finance teams believe a slow close is caused by too much data or too few people. In my experience it’s neither. It’s almost always caused by waiting — waiting on a signature, a confirmation, a spreadsheet from a plant manager who’s on vacation.

One wholesale business I worked with closed their books in 11 days. Their CFO had two open roles she couldn’t fill, so she assumed the delay was headcount. We tracked every hour for one close cycle. Forty percent of the elapsed time was pure waiting, not working.

Where 11 days actually goes

Here’s the breakdown from that tracking exercise. It looked like almost every mid-size company I’ve been inside.

• Days 1–3: chasing inventory counts from three locations that reported in different formats.
• Days 4–6: reconciling intercompany transfers that had been recorded by hand.
• Days 7–8: waiting on two department heads to approve expense reports sitting in an inbox.
• Days 9–11: building the management report itself, mostly by copying numbers between files.

Only about three of those eleven days were real accounting judgment. The rest was logistics. You cannot hire your way out of a logistics problem.

The two changes that cut it to four days

The first was a hard cutoff. All inventory counts had to be submitted by noon on day one, in one template, from all three sites. Two locations complained. Then they adjusted.

The second was a rules-based approval layer. Purchases under $2,500 and expense reports under $500 auto-approved if they matched policy. Anything above that had a 24-hour clock — if the approver missed it, it escalated to their manager automatically, no reminder emails. Average approval time went from four days to nine hours.

Then they rebuilt the management report once. One template, tied to the same numbers the close produced, refreshed on its own. The report stopped being a task and became a byproduct. None of this required a new accounting system — it required someone to treat the close as a workflow, which is where most ERP Consulting engagements actually start.

What to fix first if you only have one month

1. Record where the time goes for one close cycle. Use a shared spreadsheet, not memory. Two hours total, and genuinely eye-opening.
2. Find your top three waiting points and put a deadline on each one.
3. Automate approvals below a threshold you can defend. Start at $500 if you’re nervous.
4. Kill any report built by copying between files. Rebuild it once, properly, or stop producing it.

That same CFO now closes in four days with the same two open roles. She eventually filled them and used the capacity for margin analysis instead of counting inventory.

“We billed 61% of our own hours last quarter.” The managing partner said it flatly, like he was reading a weather report. Fifty-two consultants, a solid client roster, and nearly four of every ten hours they worked never made it onto an invoice.

His first guess was that people were coasting. They weren’t. They were working plenty — just not in a way anyone could bill for.

Where the missing hours go

We pulled three months of calendars, badge data, and project notes from twelve consultants. The pattern was consistent and, honestly, boring.

• Scope creep nobody wrote down: “quick calls” that ran 45 minutes, three times a week, per client.
• Work done before a change order was signed, sometimes weeks before.
• Internal rework caused by requirements that got clarified verbally instead of in writing.
• Travel and prep time everyone assumed was baked into the fee. It wasn’t.

None of that is dishonest. It’s just invisible. Unlogged work is unbillable work, and it compounds quietly until the year-end numbers look like a math error.

The Monday morning timesheet test

Their old process was a monthly timesheet filled in from memory on the last Friday. Memory is generous with meetings and stingy with detail. We replaced it with a five-minute daily habit and one rule that made it stick.

The rule: nobody submits a timesheet for a day they can’t describe in a single sentence per client. If the sentence is vague — “worked on the project” — the project manager sends it back. Three weeks was enough for people to stop writing vague sentences.

Then we wired hours against the original estimate so a project manager could see the gap on Monday morning instead of discovering it at invoicing. The fix wasn’t software. ERP Consulting at this stage is about changing one habit and giving managers a weekly number to look at.

Turning billing into a 3-day cycle

1. Every engagement gets an estimate broken into phases, not one lump sum.
2. Any work outside the signed scope gets a change order within 48 hours — one page, a price, a signature.
3. Timesheets close Tuesday at noon. Invoices go out Wednesday and Thursday.
4. Project managers review hours versus estimate weekly and flag anything past 85% of budget.

Nine months later the realization rate was 84%, up from 61%. On their revenue base, that was about $1.1 million a year that had been walking out the door with a smile. Same clients, same headcount.

Run one number this week: total hours your team recorded last month, divided by total hours they were paid for. If it’s under 80%, you don’t have a discipline problem. You have a visibility problem.

Start with the number: $2.3 million. That was the value of inventory in one distributor’s warehouse that had not moved in 400 days or more. It occupied 19% of their rack space and generated zero revenue.

The owner knew some of it was old. He guessed $300,000. When we pulled the aging report and walked the aisles together, confirming the real figure took four hours and made him quiet for the rest of the afternoon.

How dead stock hides in plain sight

Dead inventory doesn’t announce itself. It sits on the top rack, behind the fast movers, in packaging that still looks current. Their system tracked when items came in, not when they last went out, so most of that pile showed a “last received” date nobody ever questioned.

Meanwhile, the buying team kept reordering off minimum and maximum rules nobody had touched since the business was half its current size. Twelve percent of their purchase orders were for items that had not sold a single unit in the prior year.

The 12-month rule that changed buying behavior

• Any item with zero sales in 12 months gets flagged automatically and frozen from reordering.
• Any item with zero sales in 24 months gets a decision deadline: discount, bundle, return to vendor, or write off.
• Buyers see the flag before they can approve a purchase — no override without a written reason.
• The monthly purchasing review opens with the dead-stock list, not the sales list.

That last one matters more than it sounds. Their Monday purchasing meeting used to start with what was selling. Now it starts with what isn’t. Within two quarters, dead stock dropped from $2.3 million to $740,000 without a single layoff or fire sale.

What to do with the stock you already own

You can’t flag your way out of inventory you already bought. But you can stop pretending it’s an asset. This is the unglamorous half of ERP Consulting: nobody wants to book a $310,000 write-off, yet that’s what makes the remaining numbers believable.

1. Still sellable but slow: bundled with a fast mover at a small discount. Recovered about 60 cents on the dollar.
2. Obsolete but usable: sold to a competitor in an adjacent region who still serviced that older equipment line.
3. Damaged or expired: written off in one clean entry so the balance sheet told the truth.
4. Genuinely unknown: quarantined in one aisle, priced at cost, reviewed quarterly.

The write-off freed 1,400 square feet of rack space. They canceled a warehouse expansion that would have cost $480,000. That trade alone paid for the cleanup twice over.

Run your aging report sorted by last-ship date, not last-received date. Most systems can do it. If yours can’t, that’s the first thing worth fixing, and an ERP Consulting engagement usually surfaces it in week one.

Six months ago, a metal fabricator with 140 people on the floor had the same crisis every Tuesday. The Monday night shift would start a job, discover the brackets weren’t there, and stop the line. Forty people stood around waiting for someone to drive 90 minutes to a supplier.

Their planning screen said 1,200 brackets in stock. The shelf held 304. Nobody had stolen anything. The numbers had simply drifted, one small mistake at a time, for about four years. That gap was costing them roughly $18,000 a week in idle labor and expedited freight.

The number on the screen is not the number on the shelf

Most operations managers I meet already know their inventory accuracy is bad. What they don’t know is how bad, because nobody counts. They trust the system, the system trusts the last person who typed into it, and that person was in a hurry.

We ran a two-day cycle count on 60 of their highest-moving parts. Accuracy came back at 71%. For a shop their size, 97% or better on those items is the bar. The 26-point gap was the entire production problem — not scheduling, not the shop floor, not machine capacity.

• Pick your 50 highest-value or highest-volume parts. Not all 4,000 — just 50.
• Count them by hand, twice, on two different days, by two different people.
• Compare both counts to the system. Write the variance in dollars, not units.
• Anything off by more than 3% gets a root-cause note: wrong unit of measure, unreceived delivery, scrap never reported.

What the first 90 days actually looked like

They didn’t rip out their system. They didn’t spend much money either. They fixed the four habits that were feeding bad data into it.

1. Receiving moved to the dock. Nobody walked a delivery into the warehouse and typed it in later. It got scanned at the door, against the purchase order, before the truck pulled away.
2. Scrap got a two-minute form. Operators had been tossing bad parts in a bin and moving on. Now every scrapped part hits a tablet, and the bin gets emptied at shift end.
3. Substitutes needed approval. When a buyer swapped a similar part to keep a job moving, the swap got flagged the same day instead of never.
4. Fridays became count day for the top 50. Thirty minutes, four people, a rotating list.

By day 90, accuracy on those 50 parts hit 98.6%. Tuesday stoppages went from three a month to zero for eleven straight weeks. That’s the kind of result ERP Consulting is actually paid for — not the software, the habits wrapped around it.

The part that usually gets skipped

Pick the 50 parts this week. Count them Friday afternoon. You’ll have your answer by 4pm, and it will probably be uncomfortable.

The uncomfortable number is useful. It gives you a dollar figure to put in front of ownership, and it turns “our data is bad” from a complaint into a budget line. Do that before you call an ERP Consulting firm or sign anything.