CASE STUDY: The Obligation Map: A Simple Tool to Track Every Contractual Commitment You've Made
Updated: Sep 20
19 agreements. 147 obligations. Three hours and ten minutes. Four of them were already being breached, and the most expensive finding was not a breach at all.
A model you can audit is worth more than a testimonial you cannot.
The short version
The business: 58-person chilled ready meals manufacturer, one production site, annual revenue $14.2m
The portfolio: 19 live agreements. Two grocery supply contracts, one foodservice distribution agreement, packaging, ingredients, logistics, equipment, software
What triggered it: a customer assurance questionnaire asking for one certificate
Time to map all 19: 3 hours 10 minutes, over two afternoons. Exactly 10 minutes per contract
Obligations found: 147, an average of just under eight per agreement
Obligations with no named owner: 12
Already being breached that day: 4
Would have been breached within three weeks: 1
Net cash effect, first 12 months: $109,806.00
Which works out at: $34,675 for every hour spent building the map
She was not running a sloppy business. She had a contract folder, a renewal calendar and an auditor who had never raised a finding. What she did not have was a list of the promises the business was currently making.
Part 1. How it surfaced
She is the finance director. Fifty-eight staff, one production site, chilled ready meals into two national grocery chains and, through a distributor, into contract catering.
On 1 May 2026 she signed off the annual insurance renewal. The broker had found a saving. Product recall cover came down from $5m to $2m, and the premium came down by $4,140 a year. She approved it in under a minute, because it looked exactly like the sort of small, sensible decision a finance director is paid to make quickly.
On 11 June the larger grocery customer sent its annual supplier assurance questionnaire. Twenty-two questions, mostly routine. Question 14 asked for a current certificate of insurance showing the product recall limit.
She pulled the supply agreement for the first time in two years, and found this in clause 11.2:
The Supplier shall maintain throughout the Term product recall insurance with a limit of indemnity of not less than $5,000,000 in respect of any one claim and products liability insurance with a limit of not less than $10,000,000, and shall on request provide certificates evidencing such cover.
The business had been in breach of its largest customer contract for forty-two days, and the breach had been created by a cost saving she had personally approved.
Nothing had gone wrong. No claim, no recall, no complaint. The only thing that happened was that somebody asked.
That is the question that started the map: what else have we promised that nobody is watching?
Part 2. What an Obligation Map is, and what it is not
An Obligation Map is one row for every promise your business has made in a contract, with a name against it.
That is the whole idea. It sounds too simple to be worth three hours, and the reason it is worth three hours is that almost nobody has one.
Most businesses track contracts three ways, and all three miss the same thing.
The contract folder tells you the agreement exists. It does not tell you what is in it.
The renewal calendar tells you when the agreement ends. It is genuinely useful, and it is one row of the map, not the map.
The approvals process tells you the agreement was signed by somebody senior. It says nothing about the twenty years of Tuesdays between signature and expiry, when the promises actually have to be kept.
An Obligation Map is different in one specific way: it is organised by promise, not by document. It answers a question your filing system cannot, which is what is this business obliged to do this month, and who is doing it.
Be clear about what it is not. It is not a contract review, so it does not tell you whether the agreement is fair, whether the liability cap is sensible or whether the indemnity is one-sided. Those are real questions and they take longer. It is not a compliance audit either. It records what your contracts require. Whatever your sector rules, food safety standards or data protection duties require is a separate question and a separate list, and the two lists overlap far less than people assume.
The map does one job. It converts a filing cabinet into a to-do list.
Part 3. HANDS: the five types of obligation
Every obligation in a commercial contract falls into one of five types. Once you know the five, you stop reading and start hunting, and ten minutes is enough.
The five spell HANDS.
H is for HOLD. A state you must maintain continuously. Insurance limits, certifications, accreditations, licences, minimum stock cover, security standards, staffing levels.
Search for: at all times · throughout the Term · maintain · keep in force · not less than · insurance · accredit · certif · licence · standard
A is for ACT. Something you must actively do, usually on a repeating clock. Reports, audits, inspections, service intervals, training, mock recalls, review meetings.
Search for: shall provide · shall deliver · shall permit · report · records · audit · inspection · training · service · test · annually · quarterly · each
N is for NOTIFY. Something you must tell them, usually inside a window. Change of specification, supplier or site, change of control, incidents, claims, insolvency, notice of non-renewal.
Search for: notify · inform · give notice · in writing · within · as soon as · promptly · without delay · change of · change in
D is for DON'T. A restriction on what you may do. Exclusivity, non-solicit, non-compete, assignment, subcontracting, publicity, confidentiality, most favoured pricing.
Search for: shall not · must not · may not · without the prior written consent · exclusive · sole · refrain · solicit · compete · assign · subcontract · publicity
S is for SPEND. Money you have committed to move whether or not you want to. Minimum volumes, minimum spend, take or pay, committed licences, uplifts, deposits, retentions, interest.
Search for: minimum · committed · take or pay · shortfall · at least · per annum · uplift · index · CPI · deposit · retention · interest
Two things are worth noticing about that list.
Renewal dates are not one of the five. A renewal is a Notify obligation: you must tell them, in writing, by a date, or the contract renews itself. It is one row. An important row, and only one.
The five are not equally visible. Notify and Spend obligations tend to have dates and numbers attached, so they get noticed eventually, usually by an invoice. Hold obligations have no event attached to them at all. Nothing happens on the day you fall below your insurance limit. That is precisely why they are the ones you breach.
Part 4. The build: ten minutes a contract
She mapped nineteen agreements in 190 minutes. Not because she is fast, but because the method is four searches and a decision, and it does not allow reading.
0:00-1:30 DOCUMENT SET. List every document this agreement pulls in by reference, and mark the ones you do not physically have. You cannot map what you have not been sent.
1:30-4:00 THE VERBS. Search shall, must, will, undertakes, agrees to, is responsible for, at its own cost. Every hit that lands on you is a candidate row.
4:00-6:00 THE CLOCKS. Search within, no later than, prior to, days, months, annually, quarterly, each, throughout, at all times. This separates one-off obligations from recurring ones.
6:00-8:00 THE CONSEQUENCES. Search shortfall, minimum, liquidated, charge, indemnify, terminate, suspend, remedy, cure, entitled to. Any obligation with a number attached to failure goes to the top.
8:00-9:00 THE RESTRICTIONS. Search shall not, may not, exclusive, without the prior written consent. Short pass, high value.
9:00-10:00 ASSIGN. Every row gets a HANDS letter, an owner, a clock and a status. No exceptions.
Each row on the map has seven fields:
Contract
Clause
The obligation, in one plain sentence that a person who has never read the contract could act on
Type: H, A, N, D or S
Clock: one-off date, recurring interval, or continuous
Owner: a named person, never a department
Consequence if missed, quoted from the clause, in money wherever the contract puts a number on it
And a status. Green: owned, evidenced, in date. Amber: no owner, or you cannot evidence it. Red: breaching today, or a consequence with a number attached and nobody watching.
The one rule that makes the map work is field 6. An obligation with a department in the owner column is not tracked. It is written down, which is a different and much less useful thing. "Operations" has never once renewed a certificate.
Part 5. The five findings
Her map produced 147 rows. Twelve had nobody's name against them. Four were red on the day she built it. One turned red three weeks later, and that one was the interesting case.
Here is one from each letter.
H is for HOLD. The insurance limit that fell below the contract.
The finding that started everything, and the cheapest one to describe.
Clause 11.2 required $5,000,000 of product recall cover throughout the Term. On 1 May the limit became $2,000,000. Nothing alerted anybody, because nothing was scheduled to happen. A Hold obligation has no event attached to it.
Clause 11.5 mattered more than 11.2:
If the Supplier fails to maintain the insurances required by clause 11.2, the Customer may terminate this Agreement on thirty (30) days' written notice without liability.
The agreement carried $2,340,000 a year of revenue. The saving that put it at risk was $4,140 a year.
She restored the $5m limit on 16 June: premium back up by $4,140, plus a $310 mid-term adjustment fee. Cost to fix: $4,450. The certificate went back with the questionnaire on 19 June and the customer never knew there had been a gap.
Straightforward Spend view. Insurance covenants are the single most commonly breached Hold obligation we see in supply contracts, and the reason is structural rather than careless. Insurance is renewed by finance, on an annual cycle, against a premium. The requirement lives in a contract signed by sales, on a different cycle, against a customer. The two documents never meet unless somebody makes them meet. Put every contractual insurance minimum on one page and give it to your broker before the renewal, not after.
S is for SPEND. The minimum volume nobody was counting.
This was the biggest number on the map, and it is worth sitting with, because it was not a breach. Nothing was going wrong. The contract was working exactly as written.
Packaging supply agreement, clause 4.1:
The Customer shall purchase a Minimum Annual Volume of 4,200,000 Units in each Contract Year. Where actual purchases in a Contract Year fall below the Minimum Annual Volume, the Customer shall pay a shortfall charge calculated as the difference multiplied by the Shortfall Rate of $0.14 per Unit, invoiced within thirty (30) days of the end of that Contract Year.
The Contract Year runs from 1 January. She pulled the purchase ledger to 31 May.
Units purchased to 31 May | 1,479,000 |
Run rate | 1,479,000 ÷ 5 = 295,800 per month |
Full year projection | 295,800 × 12 = 3,549,600 |
Minimum Annual Volume | 4,200,000 |
Projected shortfall | 650,400 units |
Shortfall charge at $0.14 | $91,056.00 |
A $91,056.00 invoice, arriving in January, for units nobody had failed to order. There was no error anywhere. Volumes had drifted down when a slow line was rationalised eighteen months earlier, and the commitment had never followed them down.
Two moves, both only available because it was June and not January.
First, she consolidated. A second product line was buying its trays and film from a different supplier at a marginally better unit price. Moving it across the remaining seven months put 384,000 units through the committed agreement.
Second, she negotiated. In exchange for extending the term by twelve months, the Minimum Annual Volume came down from 4,200,000 to 3,900,000 for this Contract Year and next, with a CPI-linked cap of 3% on unit price.
Projected volume 3,549,600 + 384,000 = 3,933,600Revised minimum 3,900,000Headroom 33,600 unitsShortfall charge $0.00
$91,056.00 avoided. In January the same finding would have been an invoice with nothing to discuss.
N is for NOTIFY. The ingredient change that needed thirty days' notice and got none.
Grocery Supply Agreement, clause 7.4:
The Supplier shall give the Customer not less than thirty (30) days' prior written notice of, and shall obtain the Customer's prior written approval to, any change to the Specification, recipe, raw material supplier or manufacturing site of any Product.
And clause 7.6:
Where a change is made without the approval required by clause 7.4, the Customer shall be entitled to charge a specification variation recovery charge of $12,500 per affected Product line and to de-list any affected Product on fourteen (14) days' written notice.
In March, procurement had switched passata supplier after a price rise. Same specification, same certification, better price, three affected SKUs. It was a good decision made by a competent buyer who had never been shown clause 7.4, because clause 7.4 lived in a contract that sales had signed.
Recovery charge 3 SKUs × $12,500 = $37,500Revenue on the three SKUs at de-listing risk = $410,000 a year
She did not wait to be found. On 22 June she disclosed voluntarily with a complete specification pack: the new supplier's certification, comparative analysis, allergen statement and traceability documentation. Approval came back retrospectively on 9 July with no charge levied and written confirmation on file.
The approval was not guaranteed and she knew it. But a voluntary disclosure with a complete pack is a fundamentally different conversation to the same fact discovered at a customer audit eight months later, and the difference is worth $37,500 and three product lines.
A is for ACT. Two obligations that had never once been performed.
Act obligations are the ones people find most surprising, because they are things the business genuinely intended to do and simply never diarised.
The warranty condition. A continuous-flow chiller bought in March 2024 for $148,000, with a five-year warranty. Clause 8.3:
The warranty in clause 8.1 is conditional upon the Equipment being serviced at intervals of not more than six (6) months by an engineer approved in writing by the Supplier, and upon service records being retained and produced on request.
The last four services had been done by their own maintenance fitter, who is excellent and is not an approved engineer. The warranty was void, with 32 months left to run, on a machine whose compressor and control module the supplier quotes at $23,400 to replace.
Fix: joined the approved service plan at $3,200 a year, with written confirmation that cover resumes from the next scheduled service.
The report that was never sent. Foodservice Distribution Agreement, clause 9.2, required a quarterly volume and territory report within twenty business days of quarter end. It had never been sent. Not once in twenty-six months, which is ten missed reports.
Nobody chased it, because the report existed for the distributor's benefit. But clause 9.4 said that where the Supplier fails to provide the reports required by clause 9.2, the annual rebate reconciliation shall be calculated on the Distributor's records. The distributor's records produced a rebate claim $6,900 higher than the production ledger did.
Reports reinstated, historic quarters submitted, reconciliation agreed on the correct figures. $6,900 held.
D is for DON'T. The clause that would have made next month's deal a breach.
This is the one that turned red three weeks after the map was built, and it is the strongest argument for doing this before you need to rather than after.
Foodservice Distribution Agreement, clause 5.1:
The Supplier appoints the Distributor as its exclusive distributor of the Products to the Foodservice Channel within the Territory and shall not, during the Term, whether directly or indirectly, supply or offer to supply Products to any Foodservice Customer.
"Foodservice Customer" was defined in Schedule 2, eleven pages later, to include contract caterers.
Sales were three weeks from signing a direct supply deal with a national contract caterer. $640,000 a year, and at a 26% contribution margin, $166,400 of contribution. Everybody in the business regarded it as unambiguously good news.
Clause 5.4:
In the event of any supply made in breach of clause 5.1, the Supplier shall pay to the Distributor liquidated damages equal to fifteen per cent (15%) of the invoiced value of such supply, and the Distributor may terminate this Agreement forthwith.
Liquidated damages on a year of that deal $640,000 × 15% = $96,000Distribution agreement revenue exposed to termination = $1,900,000 a year
The map found it on 16 June, before signature rather than after. That timing is the entire value, because it turned a breach into a negotiation.
She went to the distributor first, openly. The outcome, agreed on 4 August: a written carve-out for three named national accounts, in exchange for extending the exclusivity term by twelve months and funding $18,000 of joint trade marketing.
The deal was signed on 14 August. Legally, at full margin, with the distributor as a supporter rather than a claimant.
Signed three weeks earlier and unread, it would have been a 15% liquidated damages claim, a terminated distribution agreement and a $1.9m hole where the foodservice channel used to be.
Part 6. The math, in full
Three columns, and they are kept apart deliberately, because mixing them is how contract content becomes marketing.
Column one. Charges avoided in the first twelve months. Money that would have been invoiced.
S. Volume shortfall charge, clause 4.1. 650,400 units × $0.14 = $91,056.00
N. Specification recovery charge, clause 7.6. 3 SKUs × $12,500 = $37,500.00
A. Rebate reconciliation held at the correct figure, clause 9.4 = $6,900.00
Subtotal = $135,456.00
Column two. What it cost to put the findings right.
H. Product recall limit restored to $5m: premium $4,140 plus $310 mid-term adjustment = ($4,450.00)
A. Approved service plan, year one = ($3,200.00)
D. Joint trade marketing agreed in exchange for the exclusivity carve-out = ($18,000.00)
Subtotal = ($25,650.00)
Net cash effect in twelve months.
Charges avoided $135,456.00Cost of the fixes ($25,650.00)Net $109,806.00
Time invested 3 hours 10 minutes = 190 minutesRate $109,806 × 60 divided by 190 = $34,675 an hour
Column three. Exposure closed, and deliberately not added to the number above.
These are not savings. They are risks that stopped being live, and putting them in a headline total is exactly the sort of arithmetic this business exists to argue against.
Uninsured product recall gap, closed: $3,000,000
Liquidated damages on a year of breaching supply: $96,000
Warranty reinstated on a quoted failure mode: $23,400
Annual revenue under the two agreements in breach: $4,240,000
Annual revenue on three product lines exposed to de-listing: $410,000
New business signed legally rather than in breach: $640,000 a year
Notice which column the largest number is in. That is normal, and it is the reason the map is worth building even in a year when it finds no invoice at all.
Part 7. Why this happens to well run businesses
She runs a good finance function. The business passes its audits. It still had four live breaches and a fifth three weeks away.
Three mechanisms, and they compound.
Contracts are signed by one function and performed by five
A supply agreement is negotiated by sales, signed by a director, and then handed to nobody.
Its insurance covenant has to be kept by finance. Its specification notice clause has to be kept by procurement. Its service level has to be kept by operations. Its reporting clause has to be kept by whoever owns the data. Its exclusivity has to be respected by the same sales team that signed it, eighteen months later, when a different opportunity arrives.
Not one of those five people has read the document. They are not being negligent. Nobody ever gave it to them, and if somebody had, it would have been forty pages with no indication of which two lines were theirs.
The Obligation Map is the translation layer. It takes one document that nobody reads and turns it into five short lists that individual people can actually act on.
Dates get tracked. States do not.
A renewal date is an event, so a calendar can hold it. "Maintain $5m of product recall cover at all times" is a condition, and there is no such thing as a calendar entry for a condition.
Hold obligations therefore have exactly one alert mechanism, which is the other side asking. That alert arrives at the worst possible moment, because the only reasons a counterparty asks are an audit, a claim, or a dispute.
The fix is to give every Hold obligation an artificial event. Pick a date, put a review in the calendar, and make the review produce a piece of evidence: the certificate, the accreditation, the record. If a Hold obligation cannot be evidenced in under five minutes, treat it as amber whatever you believe to be true.
Breach is silent until it is expensive
Every other failure in a business announces itself. An unpaid invoice generates a reminder. A missed delivery generates a phone call. A failed batch generates an alarm.
A breached contractual obligation generates nothing at all. There is no overdue notice. There is no counter ticking anywhere you can see it. The business carries on exactly as it did the day before, and the only difference is that the other side now holds a right it did not hold last week.
Which is why the day you find out is almost never the day it started. The insurance breach was forty-two days old. The reporting failure was twenty-six months old. The warranty had been void for something over a year.
Part 8. Three rules that keep the map alive
Most Obligation Maps get built once, are genuinely useful, and are out of date within a quarter. Three rules stop that.
Rule one. Every row has a person's name, not a function. This is the difference between a map and a list. If you cannot put a name in the owner column, the row is amber by definition, and amber rows are the agenda for the next review.
Rule two. Add the map before you sign, not after. The map takes ten minutes when you already have the document open and the deal in your head. It takes forty when you come back to it in eighteen months with no memory of what was negotiated. Make "obligations mapped" the last step of signing, alongside filing the executed copy.
Rule three. Review it quarterly, and review it on change. Quarterly takes twenty minutes once the map exists, because you are checking statuses rather than rebuilding rows. But the reviews that actually catch things are the ones triggered by change: a new insurance renewal, a new supplier, a new site, a new product line, a new sales channel, a change of ownership.
Her breach was created by an insurance renewal. Her near miss was created by a new sales channel. Neither had anything to do with the contracts themselves, which had not changed at all. The map does not go out of date because contracts change. It goes out of date because your business changes underneath them.
Part 9. Run it on your own contracts this week
We have built the Obligation Map as a download, free and ungated, in two forms.
The Obligation Map worksheet (PDF). Print it. The HANDS search guide, a worked example, and the map grid itself, designed to be filled in while the document is open on screen. One sheet per contract.
The Obligation Register (spreadsheet). The same seven fields across your whole portfolio, with a dashboard that counts your red rows, your unowned rows and your next thirty days of obligations, and a worked example tab carrying the full case study above so you can change an input and watch the answer move.
Where to start
Not with all nineteen. Start with three, in this order.
Your largest customer contract. Not your largest supplier contract. Customer agreements carry the Hold obligations, and Hold obligations are the ones you are most likely to be breaching right now without knowing it.
Anything with the word "exclusive" in it. Distribution, agency, reseller, channel and referral agreements. These are the ones where a good decision made elsewhere in the business becomes a breach, and it happens with no warning and no bad intent.
Anything with a minimum, a commitment or a take or pay. Packaging, ingredients, logistics, telecoms, energy, software seats. Then compare the minimum against your actual run rate. That comparison takes four minutes and it is the single highest-yield thing in this article.
If the map turns up something you are already breaching, that is not a failure of the exercise. That is the exercise working. You now know, on a day of your choosing, which is a vastly better day to know than the one the other side picks.
The one line to take away
A contract is not a document you keep. It is a list of promises you are currently making. If nobody in the business can name them, you are not managing them, you are hoping.
Go and map one contract. Ten minutes. Start with the customer who could hurt you most.
Download: Obligation Map Template
Comment ‘Template’ on our Instagram or Facebook post, or get in touch, if you would like us to send you the Obligations Register template (Excel).
Straightforward Spend turns contracts, agreements and business cases into decisions you can trust, without the hours of reading using automated tools. Decide fast. Decide right.
We use an AI supported contract review tool to perform our first pass checks in minutes rather than manually. It expands on the Obligation Map, saving time and ensuring consistent output. We will cover that tool in the next series as well.
Straightforward Spend is contract analysis education and tools, not a legal firm. Not legal advice. Contract law and drafting practice vary by country and by state or province; take professional advice on your own agreements.
Comments