Procurement Playbook

Executive summary

A procurement playbook is a repeatable method for finding, capturing, and keeping savings on third-party spend. The core claim of this one is simple: savings are decided by the quality of your spend data and the shape of your contract calendar, not by how hard anyone negotiates. If you cannot say what you buy, from whom, at what price, and when each agreement ends, you will negotiate the wrong categories at the wrong time and book savings that never reach the P&L.

The size of the prize has a formula. For each category, savings equal the spend you can actually move in the next twelve months multiplied by a price or volume improvement you have evidence for, minus the cost to switch or to administer the change. The first term depends on the spend cube (what is the real spend and how is it classified). The second depends on the contract calendar (which spend is open for renegotiation, and when). Only the improvement rate involves negotiating skill, and you should measure it from quotes, not assume it.

The first move is not a sourcing event. It is a four-week data job: extract twelve months of accounts payable lines, clean the vendor names, classify to a category tree, and lay every contract end date next to the spend it covers.

When to use this

Use this playbook when:

• A sponsor or board has asked for a cost program and procurement is expected to deliver part of it.
• Gross margin or EBITDA is under pressure from input costs, freight, services, or software, and nobody can say which suppliers are driving it.
• Spend is spread across many vendors, sites, or acquired businesses that each buy separately.
• You inherited a company after an acquisition and have not yet seen a consolidated view of spend.
• Budget owners sign agreements and renew them without finance or operations seeing the terms.
• You are preparing for a sale and want to show that purchasing is controlled, not ad hoc.

Do not start with this playbook if the accounts payable ledger is not reconciled to the general ledger, or if invoices are coded to a handful of catch-all accounts with no line detail. Fix the data first, using the cube method in Spend Analysis: How to Build a Clean Spend Cube. Also do not use it as a substitute for a legal review: supplier agreements bind you, and you need counsel and your own contract documents before you act on any term discussed here.

Diagnostic questions

Answer these from the ERP, the AP subledger, and the contract files. Most take a day or less once the extract exists.

1. Can you produce twelve months of AP and purchase card lines at the invoice-line level, with vendor, amount, date, GL account, and cost center?
2. What share of spend, by dollars, is classified to a real category rather than "miscellaneous," "other," or a blank?
3. How many vendor records exist, and how many remain after merging duplicates and parent-child relationships?
4. What share of spend is with the ten largest suppliers, once name variants are merged?
5. For each of those ten, is there a signed agreement, what are its pricing mechanics, and when does it end or auto-renew?
6. How much spend has no agreement at all and runs on purchase orders or invoices alone?
7. Who is allowed to commit the company to a purchase, up to what amount, and is that written down?
8. Does the same item or service get bought from more than one supplier at different prices across sites or business units?
9. How many contracts auto-renew, and what is the notice window for each?
10. When a supplier raises prices, does anyone compare the invoice to the contract or the last purchase price?
11. How are savings currently reported, and does finance agree to the baseline they are measured against?
12. Who owns each of the top categories by name, and do they have a plan for the next twelve months?

How to read the answers: if questions 1 to 3 are weak, your first job is data, and any savings estimate is a guess. If question 5 or 9 is weak, you are negotiating without knowing when you have leverage, and some of your best chances will lapse inside a renewal notice window. If question 7 is weak, savings will leak back as unmanaged spend however good the negotiation was.

The framework: two assets, one decision

Procurement results come from two assets and one decision loop.

The decision loop is: build the cube, overlay the calendar, rank categories by movable spend, choose the lever for each, run the event, and track realized savings against the baseline. Then refresh the cube and start again.

The six levers

Almost every procurement initiative is one of these. Naming the lever keeps the plan honest about where the savings come from.

Compliance is the one most often ignored. A contract that gives you a negotiated price saves nothing if buyers order off contract or invoices are paid without price checks.

Step-by-step process

Each step has an owner and an output. Time-boxes assume a mid-market company with one or two ERPs.

Phase 1: Build the data (weeks 1-4)

1. Extract twelve months of AP invoice lines, purchase card transactions, and expense-reported supplier spend. Include vendor ID, vendor name, invoice date, amount, GL account, cost center, PO number if any, and unit and quantity where they exist. Owner: Controller. Output: raw extract, reconciled to the GL by total.
2. Normalize vendors. Merge name variants and remit-to records, then roll subsidiaries up to a parent. Owner: FP&A with a procurement analyst. Output: vendor master with a parent ID.
3. Classify every line to a category tree of two or three levels. Start with a standard tree, not one you design from scratch. Classify by vendor first, then override at line level where one vendor sells several categories. Owner: Procurement. Output: cube with a category on every line.
4. Measure the quality of the cube: percent of dollars classified, percent of vendors merged, percent of lines with a usable unit price. Publish these three numbers with every output. Owner: FP&A. Output: data quality scorecard.
5. Collect the contracts. Pull every agreement over a threshold you set (for example, annual value above a stated amount, or any auto-renewing term) from legal files, shared drives, and email. Owner: Procurement with Legal. Output: contract register.
6. Build the calendar. For each contract record the counterparty (matched to the parent ID), start and end dates, renewal type, notice window, pricing mechanism, minimum commitments, and termination terms as written. Match each to the spend it covers. Owner: Procurement. Output: contract calendar by month.

Phase 2: Size and prioritize (weeks 4-6)

7. Compute movable spend by category: spend with no agreement, plus spend whose agreement ends or can be terminated within the planning window. Owner: FP&A. Output: movable spend table.
8. Write the hypothesis for each category: lever, evidence for the improvement rate (a recent quote, an index, a price paid at another site), and one-time cost to switch. Owner: Category owner. Output: category one-pager.
9. Rank categories by risk-adjusted savings and effort. Owner: Procurement lead. Output: a sourcing roadmap sequenced against the calendar, so that the earliest end dates and notice deadlines drive the order.
10. Agree the baseline with finance before any event: which period, which price, which volume. Owner: CFO. Output: signed-off baseline definition.

Phase 3: Execute (weeks 6-20)

11. Prepare each negotiation or event: demand, specification, supplier list, walk-away position, and the ask. See Supplier Negotiation Prep Pack. Owner: Category owner.
12. Run the sourcing event where competition is the lever, using a documented process for specifications, bid comparison, and award. See Sourcing Event Process: RFP and RFQ. Owner: Procurement lead with the budget owner.
13. Award and contract, with legal review of terms by counsel, notice and renewal clauses diarized, and pricing loaded where buyers and AP can see it. Owner: Procurement and Legal.
14. Consolidate where several suppliers do the same job, using the method in Vendor Consolidation Playbook. Owner: Category owner.

Phase 4: Hold the savings (ongoing)

15. Load prices from the new agreements into the ordering system or an invoice check, so off-contract buying and price drift show up. Owner: AP and Procurement.
16. Report the savings ledger monthly: committed, realized, and leaked. Owner: Finance. Output: the one savings figure the CFO will defend.
17. Refresh the cube quarterly and the calendar monthly. Owner: Procurement. Output: next quarter's sourcing roadmap.

Why the contract calendar decides when savings are available

Teams often rank categories by size alone. That ignores timing. A large category locked into an agreement for another thirty months is not available this year, except through the levers that do not depend on timing: specification, demand, and compliance. A smaller category with a contract that auto-renews in ninety days, with a notice window closing in forty, may be the best opportunity on the list.

Three calendar facts matter for each agreement:

• End date and renewal type. Fixed term, evergreen, or auto-renewing. Auto-renewal with a notice window is the single most common way leverage disappears: the window closes, the term rolls, and the next opportunity is a year away.
• Notice window. The latest date on which you can give notice. Diary it with a reminder well before, and name a person who is responsible.
• Commitments and exit costs. Minimum volumes, take-or-pay terms, early termination fees. These set the real cost of switching and belong in the savings calculation.

Read these as written in your own agreements, and ask counsel to confirm how any clause applies before you rely on it. This playbook describes a process, not a legal interpretation.

Then add a second dimension: supply market timing. If the market for a category has softened or a new supplier has entered, a window is more valuable than it looks. If it is tight, the right move may be to extend early on stable terms. That judgment belongs in the category one-pager, supported by quotes.

Templates

Spend cube field list (copy into a spreadsheet)

Contract calendar (copy into a spreadsheet)

Category one-pager

Savings ledger

Keep committed, realized, and leakage in separate columns. A single savings column mixes promises and results.

Metrics and KPIs

Setting targets without invented benchmarks. Use the following, in this order: your own trailing best (the lowest unit prices you have paid in the last two years, per item or rate card), contractual floors and caps (index-linked pricing, volume tiers you already have), and named public sources such as published commodity indices or listed peers' disclosures where relevant. Set savings targets category by category from quotes and the movable-spend table, and set data targets as a trajectory: for example, classified spend rising each quarter against the figure you measured at baseline. Do not import a "typical savings percentage" from an article and apply it to your categories.

Common mistakes

1. Negotiating before the data is clean. Symptom: the largest supplier by name turns out to be three records for one vendor, or a category total moves by a large amount between two presentations. Fix the cube first, and publish its quality numbers.
2. Ranking by size alone. Symptom: the roadmap starts with the biggest category, which is locked in for years, while a smaller agreement renews unnoticed. Rank by movable spend.
3. Missing the notice window. Symptom: the first the company hears of a renewal is the new invoice. Diary every notice date and name an owner.
4. Counting savings against a soft baseline. Symptom: savings are measured against a budget or a list price rather than the last price paid, and finance cannot trace them. Agree the baseline in writing before the event.
5. Declaring savings at award. Symptom: committed savings are reported as results. Track committed and realized separately and reconcile to the P&L.
6. Ignoring compliance. Symptom: contract prices exist, but buyers still order from the old supplier. Load prices where purchasing happens and check invoices against them.
7. Cutting suppliers without a continuity view. Symptom: a single source for a critical input, with no qualified backup. Consolidation needs a concentration limit and a fallback.
8. Treating procurement as an annual project. Symptom: savings fade by the next budget cycle. The cube, calendar, and ledger are operating records, not deliverables.
9. Leaving the budget owner out. Symptom: the new supplier is technically cheaper but the user refuses it. Include the user of the spend in the specification and the decision.
10. Using legal terms from memory. Symptom: someone assumes a contract can be exited at will. Read the actual clause and involve counsel before you commit to a plan that depends on it.

Worked example

Illustrative. The numbers below are invented to show the method. They are not benchmarks, and the improvement rates are assumptions you would replace with your own quotes.

A manufacturer has $40.0M of third-party spend. $13.5M is raw materials, which are managed under a separate direct-materials process and excluded here. The remaining $26.5M is the addressable base for this program.

The twelve-month extract had $5.83M (22% of the $26.5M) in "other" or blank categories. After vendor merging and line-level classification, the unclassified share fell to $1.06M (4%). That cleanup moved dollars into named categories, and the category owners finally had totals they could plan against.

Reconciliation: the category spend sums to $26.5M, the movable spend sums to $16.8M (63% of the addressable base), and the hypotheses sum to $563K. The savings column is movable spend times the assumed improvement, not total spend times a rate. For packaging, $3.6M x 3% = $108K.

What the table tells the operator:

• If someone applied a single 3% rate to the full $26.5M, they would claim $795K. The calendar says only $16.8M can move this year, so that figure overstates the prize by about $232K at the same rate. Using category-specific rates and the movable base gives a smaller, defensible number.
• MRO has the largest hypothesis ($175K) although it is only the fourth largest category, because all of its spend is movable and fragmented. It is the right first event.
• Packaging is large but only 60% of it is open. The remaining $2.4M is a reminder on the calendar, with its notice date named, not an event this year.
• Professional services is $3.0M but only $1.2M is open. Specification and demand (fewer engagements, standard rate cards) may do more than a rate fight, and that is a category-plan decision.

Next, the savings ledger. Suppose the MRO event is awarded with a committed annualized saving of $175K, and nine months into the year the realized savings are $98K because some buyers still order from old suppliers. Savings conversion is 98 ÷ (175 x 9/12) = 98 ÷ 131.25 = 74.7%. The gap of about $33K is leakage, and the fix is a price-load and a compliance report, not another negotiation.

AI prompt assistance

AI can speed up cleaning and drafting. It cannot see your contracts or ledger unless you provide them, and it can misclassify or misread terms. Supply the data yourself, and check each output as described.

1. Normalize vendor names.

I will paste a list of vendor names with total spend per name. Group names that likely refer to the same company (abbreviations, legal suffixes, remit-to variants, typos) and propose one parent name per group. Return a table with the original name, proposed parent, and a confidence label. Do not merge anything you are not confident about; list those separately.

Check: review every merged group with spend above a threshold you set, and confirm with the AP team or a registration lookup. Total spend before and after merging must match exactly.

2. Classify spend lines.

Here is a category tree (pasted) and 200 invoice lines with vendor, description, and amount. Assign each line to the lowest-level category in the tree. If a line cannot be classified from the description, label it "unclassified" rather than guessing. Return a table and a count by category.

Check: sample at least 30 lines across categories, compare against what the budget owner would say, and recompute the dollar total by category.

3. Extract contract terms into the calendar.

I will paste the text of a supplier agreement. Extract: parties, effective date, term, renewal mechanism, notice period for non-renewal, pricing mechanism, minimum commitments, and termination clauses. Quote the exact clause text for each item and tell me if any item is not stated. Do not interpret legal effect.

Check: read each quoted clause yourself and have counsel confirm anything that affects an exit or notice decision. Never rely on the extraction alone for a deadline.

4. Draft a category one-pager.

Using this cube summary for the category (pasted: spend, vendor count, top-five share, contract end dates), draft a one-page category plan with demand drivers, likely levers, risks, and open questions I should answer before sourcing. List any assumptions you made.

Check: replace every assumption with a fact from your data, and confirm every figure against the cube.

Implementation guide

Governance: the CFO owns the savings target and the baseline, the procurement lead owns the cube, the calendar, and the roadmap, category owners own plans and events, Legal owns contract review, and Finance owns the savings ledger. Budget owners sign the specification and the award for spend they use. Report spend under contract, notice deadlines missed, savings conversion, and tail spend share monthly to the CEO or sponsor alongside cash.

What to expect: the data work often finds savings before any event, from duplicate payments, off-contract prices, and agreements nobody knew were running. The first events show results in the following quarter. Savings that depend on consolidation and demand change take longer and require operations to change behavior. Do not announce an annual savings figure until the baseline, the movable spend table, and the first quotes are in hand.

Downloadable assets

The spend cube field list, contract calendar, category one-pager, and savings ledger above are designed to be copied into a spreadsheet as written. The Spend Cube Starter turns a CSV of AP lines into a category rollup and vendor concentration view. It is not published yet.

Related resources

• Spend Analysis: How to Build a Clean Spend Cube: the data asset behind every step above.
• Vendor Consolidation Playbook: when fewer suppliers is the lever, and how to avoid single-source risk.
• Contract Management for Procurement Savings: building and maintaining the contract calendar.
• Supplier Negotiation Prep Pack and Sourcing Event Process: RFP and RFQ: running the events.
• Margin Expansion Playbook: where procurement sits next to pricing and operations.
• DPO: how to extend payables without damaging suppliers: the payment terms lever, with its supplier relationship costs.

Frequently asked questions

What is a procurement playbook?

A procurement playbook is a documented method for classifying spend, finding savings, running sourcing events, and keeping the savings once won. It names owners, templates, and metrics so results do not depend on one person. The strongest ones begin with a clean spend cube and a contract calendar, because those two assets decide which categories are worth working and when.

How do you calculate procurement savings?

Savings for an initiative equal (baseline unit price minus new unit price) times actual volume, measured against a baseline finance agreed in advance. Report committed and realized savings separately, and net out one-time costs to switch. Reconcile the total to the P&L lines after volume and mix effects, because savings that cannot be traced there are not yet savings.

Why does the contract calendar matter for procurement savings?

The contract calendar shows when you have leverage. Spend under a long agreement cannot be renegotiated or sourced competitively until it ends or can be exited, and an auto-renewal notice window that passes can postpone the opportunity a full term. Sequence your sourcing events by notice deadlines, and read each agreement as written, with counsel confirming how clauses apply.

Where should a mid-market company start with procurement?

Start with a twelve-month extract of AP invoice lines, cleaned for vendor names and classified to categories, and a register of contract end dates and notice windows. That takes about a month and gives you movable spend by category. Then pick the first events where spend is large, open, and fragmented, and agree the savings baseline with finance before you begin.

How long does it take to see procurement savings?

The data work can surface savings within weeks, such as duplicate payments, off-contract prices, and unknown auto-renewals. Competitive events typically deliver the following quarter, once awards and price loads are done. Consolidation and demand reduction take longer because operations must change behavior. Track realized savings, not committed, and avoid announcing a figure before quotes are in.

How this works in a System of Performance

Procurement results decay when the cube, the contract terms, and the realized prices live in different files that nobody reconciles. A Performis routine can take on a recurring job like this, with the scope and fit agreed with you: it works from the records your team supplies, prepares the work for your named reviewer, and keeps each approval on record. To see how much of your own spend is movable and how clean your vendor data is, bring it to us as a launch partner.