For Electrical distributors
Electrical: the cost on your AP invoice is not your cost, so a floor computed from it is not your floor.
This page leads with a limitation because it decides everything after it: where purchases are claimed back through special pricing agreements, a floor computed from invoiced cost sits above your real one and will hold lines that are perfectly profitable.
A working session on your numbers, not a slide deck. No account, no card, and nothing connected until you say so.
Four things that go wrong on a electrical book
Named rather than categorized. “Margin erosion” is a heading a reader cannot check against their own week; each of these is something that happened on a Tuesday, and if none of them is familiar then this is the wrong page and the rest of it will not improve.
Problem one
The invoiced cost is the wrong number
Under ship-and-debit and claimback mechanics you buy at standard cost and claim the difference afterwards. The AP-invoice line — the only cost PIE reads, and the only cost most systems record against the purchase — is therefore higher than what the product actually cost you. Everything computed from it inherits that error in a known direction. Worse, it is wrong by a different amount per manufacturer and per agreement, so there is no single correction anybody could apply. The error has a structure, and the structure lives in a document nobody opens at quote time.
- The rebate arrives after the customer has been invoiced
- Nothing in the purchase record anticipates it
- So cost is overstated on exactly the lines that matter most
- Wrong by a different amount per manufacturer and per agreement
Problem two
Which means the failure is over-holding, not under-selling
A floor is cost divided by one minus your margin floor, so an inflated cost gives an inflated floor. On an agreement-heavy book a control built this way objects to lines that are genuinely profitable — and a control that is wrong often enough gets overridden as a matter of routine within a month. After that it is dead on the lines where it was right, which is worse than never having had it. That is a specific and unusual failure: the product does not lose you money, it loses you the control. And it does the damage fastest on your largest accounts, because those are the ones the agreements cover.
- The error is systematic, not random, and points one way
- An override habit is learned quickly and unlearned slowly
- A floor the desk ignores costs the same to run and catches nothing
- It costs you the control rather than the money
Problem three
Accounts do not resign, they reduce
A contractor prices elsewhere on one job, then on two, and the line items you used to see stop appearing while the relationship carries on looking entirely fine. Nobody reports it because nothing happened — there was no complaint, no lost tender, no conversation. By the time it is visible in a total it is a year old and the competitor is established. It is also the half of this problem your book can act on, which is why it sits here rather than in the limits: the evidence for a reduction is the invoices, and the invoices are complete.
- There is no event to report, only an absence
- A total hides it until the absence is large
- Detection needs no cost at all, which is why it works on this book
- The evidence for a reduction is the invoices, and those are complete
Problem four
Nothing in the quote knows which lines were claimable
The agreement is administered after the invoice, in a different system or a different spreadsheet, by somebody who is not the person quoting. So the person deciding a price does not know whether this line is one the agreement covers, and a legitimate claim that nobody files is margin you earned and did not collect. So the desk's own instinct about which lines are worth having is trained on the wrong number, month after month, and it becomes more confident rather than more accurate.
- The quoting desk and the claims desk are different people
- They work from different records, at different times
- PIE does not join them — see the limits below
- The desk's instinct is trained on the wrong number
What it costs is the way it accumulates
Two mechanisms pointing in opposite directions, and it matters a great deal which one your book is exposed to.
How it compounds
A quiet decline compounds the ordinary way: the conversation you eventually have is a retrieval rather than a save. The rebate exposure compounds differently and faster, because what it destroys is not margin but the control itself. Every wrong hold teaches the desk that the floor is noise. Twenty of them and the override is reflexive, and the one line a month that genuinely needed stopping goes out with all the others.
Why nothing catches it
This is the exact shape of failure the platform's own rules exist to prevent, which is why the page states it rather than selling around it: a number computed correctly from the wrong input is indistinguishable from a number computed correctly. It carries the same provenance, the same policy version and the same audit trail as a right answer.
No figure on this page is a claim about your book. There are no industry averages here and no recovered-margin percentages: the only numbers on the page are on the card further down, and they are sample figures re-derived by a test from the operands the card itself shows.
What a electrical desk does about it today
Not a list of mistakes. Every one of these is a sensible response to a real constraint, which is exactly why it has survived — and the last one is usually the most rational of them. A product that does not understand why the current answer is the current answer has nothing useful to replace it with.
A rebate spreadsheet owned by finance
Accurate, reconciled monthly or quarterly, and not available to the person quoting. It is the real cost record in the building, and it is not in the quoting flow.
The agreement letters, in a folder
The terms exist in writing, per manufacturer and per job. Finding out whether this line falls under one takes longer than quoting the line does, so mostly it is not looked up.
A margin report that is wrong in a known direction
Everybody who reads it knows to mentally add something back for rebates. That adjustment is a habit rather than a figure, it differs per reader, and it is applied to a whole-book average.
Nothing, on decline
The retention half usually has no process at all — not because it is hard, but because nothing raises it. An account that reduces produces no event for anyone to respond to.
Where PIE intervenes, and at which moment
Every figure below is deterministic arithmetic over rows your ERP already wrote, stamped with the version of the margin policy that judged it. The AI on this platform reads those numbers and phrases them; it never produces one. What is not listed here does not apply to a electrical book, and the limits further down say which of those are permanent.
Decline and dormancy, which need no cost
Four detectors run over invoice history alone — decline, dormancy, margin deterioration and cost that was never passed through — plus six at the customer-and-item grain. Each opens into the rows that raised it. The first two need no purchase cost whatsoever to be right, which is why this is the half of the product that works on your book today and the half the page leads with.
Where your price stopped following your cost
Even on an agreement-heavy book the *direction* of your invoiced cost is real: when it rose and your price did not, the pass-through detector names the customer-and-item pairs. It is the rebate-independent half of the margin question, and usually the actionable half, because the answer is a price conversation rather than a claim.
The enquiry, resolved against your own book
Paste a takeoff or an emailed enquiry and PIE reads it into lines, resolving each against the catalog you uploaded by exact match and by nearest neighbour over the description. On a book of clean manufacturer catalog numbers the exact match usually carries it, and no model is asked what a line is. It finds the line in your own book and proposes no substitute from another manufacturer's.
And not a margin floor, on this book, today
This is the paragraph that is missing on purpose. Every other trade page in this family leads with a per-line floor; on an agreement-heavy electrical book that floor is computed from a cost that is not your cost, so the platform can enforce whatever policy you set and cannot promise the policy means what you think it means. Whatever floor you do set, the sign-off on a breach is append-only and carries the version of the policy in force when it was given — so editing the policy later does not make last quarter's decisions unreadable. That is the whole claim.
One quote line, start to finish
Not a floor, on this page. The worked example is a declining account, because that is the half of the product your book can actually support. The card below is drawn from a floor decision and is shown for what it is — sample figures, and the mechanism this trade cannot yet trust.
THHN 12 STR CU 500FT
200 units for A repeat account in Ohio, quoted at $41.20 — routed to a sales manager, no individual owner. The platform holds it, not the salesperson.
| Quoted unit pricethis quote | $41.20 |
| Effective unit costbill | $38.08 |
| Margin flooryour margin policy | $44.80 |
| Recommendedthis customer’s own history | $48.70 |
| Margin at this pricecomputed | 7.6% |
The manager’s decision card, as the product draws it — with sample figures, because no customer’s numbers belong on a public page.
The rows are already there
Your invoices synced on the first pull — eighteen months by default, earlier if you set an earlier date. Nothing here needs a purchase cost, a rebate record or a price book.
A detector reads one account's history
Deterministic arithmetic against thresholds you set: this contractor's monthly line count and revenue against their own prior pattern, per item as well as in total. A model is never asked whether something is worth raising.
It raises the reduction, not the total
The account's revenue is down a little and its mix is down a lot — the items they stopped buying are named, and the signal opens into the rows that produced it rather than into a score.
Somebody has the conversation while it is a save
Months earlier than a year-end review, and armed with what stopped rather than with a feeling that something has. Whether the conversation works is your business; having it in time is the product's.
And the floor stays off, until the rebate layer exists
The card's arithmetic is real and its figures are samples. On your book the cost operand is overstated, so the floor it produces is too high and the platform would hold profitable lines. This page will not pretend otherwise, and there is no date.
Sample figures throughout. No customer’s numbers appear on a public page, every figure on the card is re-derived by a test from the same operands it shows, and the floor is cost divided by one minus the margin floor your policy sets — nothing rounded in your favor, because a floor rounded down is a floor that has moved.
The limits, before you find them in week three
Every item below is a thing this platform does not do on a electrical book today. They are on the page because a capability list is not what a distributor who has survived one implementation believes — and because the alternative is you discovering them after a pilot.
- Special pricing agreements, ship-and-debit and claim-backs are not modelled at all, and on this book that is the headline. PIE computes margin from the cost on the AP-invoice line; where that cost is later reduced by a claim, the floor PIE computes is above your real one and it will hold lines that are profitable. Treat the margin half of this product as unproven on an agreement-heavy book until the rebate layer exists.
- Rebate accrual, claim tracking and unclaimed-rebate reporting are not features here. If that is the problem you are solving, buy something built for it.
- Quotes are not imported from any ERP, so a win rate has no denominator until you start quoting here.
- Nothing resolves until a price list has been uploaded and its decoding confirmed by a person.
- Salespeople are not imported from any ERP, so every approval routes to management until accounts are assigned inside PIE.
Why it can be trusted
The AI never computes a single number.
Your business data and your rules determine the number. AI explains it.
Same inputs, same answer, every time — with a paper trail.
- Deterministic calculations — every figure is arithmetic on your own records; turn AI off and every number still works
- Auditable decisions — each number names the policy that produced it, so a price you quoted last quarter still explains itself
- Your data — read from your own books, used for you alone, exportable and erasable on request; AI runs on your own account
- Your rules — you set the floors and thresholds, and PIE holds every quote to them
- Provenance — every row carries the connector, the connection and the id it came from, so nothing in PIE claims a source it did not have
What electrical distributors ask first
Most of our cost is SPA-driven. Is this product wrong for us?
The margin half of it is, until a rebate layer exists — a floor computed from invoiced cost sits above your real floor and will hold profitable lines. The detection half is not: decline, dormancy and price-versus-cost drift are computed from invoice history and do not depend on the rebate at all. Buy it for that or wait; this page is not going to pretend otherwise.
What can PIE tell us with no reliable cost?
Which accounts are declining, which have gone dormant, which customer-and-item pairs have drifted, and what each customer has paid for each item over time. That is the same set a book with no purchase cost at all gets, and it is arithmetic over rows you already have.
Does the AI decide what to escalate?
No. The detectors are deterministic arithmetic over persisted rows against thresholds you set. A model may phrase a finding; it never produces a number and never decides one is worth raising. Turn it off and the same list appears.
Will you build the rebate layer?
It is the largest single item on the roadmap and it unlocks this trade and plumbing together. There is no date, and this page will not give you one — what it will do is tell you plainly that the floor is not trustworthy on your book today.
Which of the seven this trade runs
Prophet 21 is a core Epicor vertical here and is what a large part of this trade runs. The other systems this trade commonly runs — Eclipse, a Trade Service feed — are not among the seven PIE reads, and that is a hard gate rather than a caveat: where the book cannot be read, nothing on this page applies.
Your book already knows where the margin went.
30–45 minutes, your own numbers on the screen, and an honest answer about what PIE can and cannot see on a electrical book.