Commercial Layer Architecture for a Mid-Market B2B Distributor

TL;DR

A mid-market wholesale distributor operating multiple regional branches ran every transaction through a legacy ERP — orders, historical pricing, billing, account records — and had no commercial layer above it. Quotes lived in reps' email threads and local spreadsheets. Follow-up depended on memory. Every opportunity carried equal priority regardless of yield, so senior outside sales capacity drained into routine low-margin reorders, and branch-level pipeline reporting was a subjective roll-up disconnected from the order history sitting in the ERP. Kynera was engaged not to replace the ledger-of-record, but to build the layer it never had: a structured commercial dataset extracted from the ERP, a disciplined CRM layer above it, a scoring mechanism that ranks open opportunities by account value and qualification state, and routing rules that put expensive selling hours where they are worth spending. The result is a pipeline the business can see, rank, and staff deliberately — and a commercial record that didn't exist before, which is what any forecasting work will eventually run on.

‍

‍

The ERP knew everything except what would happen next

The distributor's ERP was doing its job. It held every completed order, every historical price, every account record, accurately and without dispute. It was the ledger-of-record and behaved like one.
‍

The problem is what a ledger-of-record is built to answer. It records what has already been sold. It holds nothing about what is being sold right now, to whom, at what stage, by which rep. That information existed — it just lived in inboxes, in local spreadsheets, and in the heads of the people carrying the accounts. No system held it together, so no one could see the pipeline as a single object.
‍

At a single branch with a handful of reps, that gap is manageable through familiarity. Across multiple regional branches, it becomes the source of nearly every commercial problem the business has.

‍

The ask

The request that reached Kynera was framed as a tooling problem: get the sales team off spreadsheets and into something structured, and give leadership visibility into the pipeline.
‍

The company had described its symptoms accurately. What it hadn't named was that these weren't separate problems — unmanaged pipeline, flat prioritization, wasted senior capacity, unreliable branch reporting — but one missing layer surfacing in several places.

‍

What was actually behind it

Each symptom traces to the same absence.
‍

Quotes fragmented across email because there was no shared commercial record to hold them. Every deal received equal attention because nothing ranked them — a routine low-margin reorder and an account with real wallet-share upside entered the queue identically, and the reps most expensive to employ absorbed whichever landed in front of them. Qualification was inconsistent because no gates were enforced, so opportunities advanced on optimism. And branch reporting relied on manager recollection precisely because the ERP's order history was never connected to anything forward-looking.
‍

That distinction shaped the solution. The instinct here is to buy a CRM platform, or to build a data warehouse first and reason about commercial process later. Both are expensive answers to a problem that didn't require either. The ERP was accurate. The gap was that nothing sat above it, translating what it already knew into how selling capacity gets allocated. The right move wasn't to replace the ledger-of-record. It was to build the commercial layer the ledger was never designed to be.

‍

‍

The approach: structure the data, then rank the pipeline, then route the capacity

Lean data foundation

Rather than commission a data platform build the business would wait months for, Kynera extracted historical account and transactional order data directly from the ERP and structured it into a clean operational commercial dataset — normalized account records, order history, pricing behavior, sector and wallet-share signals. Scope was set by what the pipeline and scoring logic actually required, not by what a general-purpose warehouse might one day need. Speed to value was the constraint, and the dataset was built against it.

‍

Commercial framework

A disciplined CRM layer was established on top of the ERP, structured around the 4R model — Reach, React, Refresh, Reengage — a model we apply when configuring CRM layers for businesses whose revenue is reorder-driven rather than one-directional. Distribution accounts don't move through a funnel once; they cycle, go dormant, and return. Standard funnel stages don't describe that motion, which is part of why generic CRM rollouts in this sector are abandoned so often.

Qualification gates were enforced at stage transitions so that opportunities advance on evidence rather than on a rep's confidence — but deliberately not at uniform weight. BANT applies across the pipeline as the baseline gate: light enough that a rep will actually complete it on a routine reorder. The heavier MEDDPICC criteria are reserved for the largest and most complex opportunities, where multiple stakeholders and a real evaluation process justify the additional discipline. Applying enterprise-grade qualification to every transaction is the fastest way to have reps stop using the system altogether.

‍

Opportunity scoring

A scoring mechanism ranks every open opportunity so the pipeline has a defensible order rather than an arbitrary one. At launch it runs on what the ERP could actually support: account fit derived from historical data — wallet share, order cadence, sector, account size — combined with the opportunity's qualification state and current engagement activity. These are rules and weights, not a trained model, and that was the correct starting point. No history of won and lost opportunities existed before this project, because no pipeline existed to record one. The scoring logic was built so that outcome data accumulates against it in a structured form from day one, which is the prerequisite for tuning the weights against real results later.

‍

Opportunity routing

Scoring only matters if it changes who does what. High-value, well-qualified opportunities route to senior outside sales capacity. Low-complexity reorders and small accounts route to inside sales or to a standardized follow-up track. The routing runs on the score rather than on availability, which addresses the two failure modes the manual process produced constantly: a senior rep spending an afternoon on a reorder that inside sales could have handled, and a valuable opportunity sitting untouched because it arrived while the right person was occupied with something smaller.

‍

Making the layer usable

A commercial layer only works if the people in it maintain the record. The system was designed to pull whatever the ERP already knows — account history, pricing, order cadence — so reps aren't re-entering data that exists elsewhere, and manual input is limited to what only a person can supply: stage, next step, qualification detail. Kynera ran working sessions with the sales and marketing teams covering how the layer operates and which fields carry weight in scoring and routing, so the discipline the system depends on is understood rather than imposed.

‍

‍

Built to extend

The engagement deliberately stopped short of building a forecasting model, and the reason is sequencing rather than scope-cutting. A forecast is only as good as the record underneath it, and until this project the distributor had no structured record of commercial activity at all — only closed transactions in the ERP and unstructured intent scattered across inboxes.
‍

What now accumulates is exactly that missing record: every opportunity scored, routed, and resolved, building a history of what was expected against what actually happened, sitting alongside known ERP demand cycles. That pairing is the input any predictive commercial model requires. A bolt-on forecasting module is currently under evaluation and sits outside this engagement's scope. The point is that when it is assessed, it starts from a foundation that already exists rather than from a data-collection project disguised as a modeling one.

‍
‍

Business impact

‍

Sales capacity

Senior selling hours moved off administration and low-yield reorder handling and onto the accounts where they are worth spending. Capacity wasn't added; it was redirected. For a distributor at this scale that distinction matters more than it sounds — outside sales headcount is among the most expensive capacity the business carries, and a meaningful share of it was being consumed by work that carried no requirement for seniority at all.

‍

Deal handling and margin discipline

The claim here is about mechanism, not conversion rates. High-value opportunities now reach an experienced rep faster and with the account's full order and pricing history attached, rather than being worked from whatever the rep could recall or reconstruct from email. Contract and historical pricing sit visible in the record at the point a quote is prepared, which removes the most common route to unapproved discounting — a rep discounting because verifying the agreed price was slower than conceding on it.

‍

Branch-level visibility

The change here is narrower than it first appears, and worth stating precisely. The business did not acquire a demand forecast. What it acquired is an objective base underneath the reporting it was already producing: a ranked, visible pipeline replacing a branch manager's recollection of which deals felt likely. Regional planning conversations now start from what is actually in play and how it is scored, rather than from an estimate assembled the week before. That is an improvement in input quality, and a precondition for genuine forecasting — not a substitute for it.

‍

‍

Key takeaway

The distributor's commercial problem was never a CRM problem. The ERP held the truth about what the business had sold and nothing about what it was about to sell — and no layer existed to turn the first into the second. Fragmented quotes, flat prioritization, misallocated senior capacity, and reporting built on recollection weren't separate issues to be fixed separately. They were one missing layer, surfacing in several places.