Mark Coulson

In their account, the benchmark becomes dangerous when its output is treated as gospel rather than a starting position.

A large client paid thousands of pounds a year for an online procurement benchmarking platform and used it on a BMC software deal.

• The tool said 15% further savings were achievable.
• Mark Coulson handed the negotiation to an IT procurement specialist with over 20 years of experience, who ignored the benchmark and secured a 44% saving, worth more than £0.5m beyond the tool's projection.

Mark Coulson does not reject benchmarking outright.

• They see value in the platforms but question whether their licence fees are justified alongside an experienced procurement professional.
• The tension is between using benchmark data as context and letting it cap the outcome.

Read original

A number spread across other operating costs cannot be defended, credited or acted on like a marketing or IT line. Their conclusion: functions taken seriously in the boardroom make their number visible before anyone asks where it is.

Mark Coulson describes sitting in on a board pack review and watching a CFO scroll for the procurement line.

• There is none.
• Procurement is folded into "other operating costs" across a dozen categories, invisible until something moves enough to prompt a question.
• Marketing has a line, and IT usually has its own slide.

Read original

In their view, most categories in most businesses are at the reactive stage. Moving up does not require adding headcount; it takes someone with the time and expertise to run the category, not just hold the title.

The post asks readers to judge which stage their biggest category would actually sit at, rather than where the organisation chart says it should. Most categories, Mark Coulson writes, are "owned on paper, invisible in practice."

Read original

The cost is not only the occasional wrong answer but the erosion of the judgement people were hired for. They offer three questions for anyone handing over AI-generated work: what did it get wrong, what would you change before a supplier saw it, and what did you know that it did not?

A procurement director described a team member pasting a question into Copilot and sending the answer straight to the director's inbox without anyone reading it.

• The director, as reported by Mark Coulson, is both frightened of the technology and convinced of its potential.
• Copilot can draft a clause in seconds, but it does not know which supplier always slips at year end or which stakeholder will quietly block the change.
• Coulson says the manager who forwards the answer is a worse problem than the person who pasted it.

The post holds two views at once: a procurement director's fear of AI and belief in its potential. Mark Coulson's position is that the tool is not the problem; accepting its first answer is.

Read original

Media owners reward agencies for the volume they place, so if rebates, bonuses and discounts earned on the client's budget are not returned, part of that budget funds someone else's margin and influences where ads run. Their practical test: ask in writing, before the pitch and before any ideas are presented, whether every rebate earned on the client's spend will come back.

Mark Coulson asked six media agencies whether they would return every rebate, bonus and discount earned on a client's money.

• Three said yes in writing, two said yes but kept margin inside "value" arrangements priced at the agency's discretion, and one incumbent declined five transparency terms outright.
• The post presents the answers as a reusable test: full transparency, qualified yes or honest no.

The claims concern media buying, not software licensing. For software contracts, ask where the supplier's margin comes from and who can see it: a headline fee may conceal the real cost.

Read original

The post gives a concrete number and challenges a common buying assumption.

Mark Coulson describes a client paying for an online procurement benchmarking platform for software SaaS.

• The platform said a further 15% saving was achievable on a BMC software deal.
• Coulson's team assigned the task to an IT procurement expert with over 20 years of experience, who set the benchmark aside and negotiated a 44% saving.
• Following the benchmark strictly would have cost the client over £0.5m.
• Coulson argues that benchmarks have value but cannot replace an experienced procurement professional who knows how far to push.

For brands planning a campaign

For a B2B SaaS company trying to cut vendor costs, the post supports an engagement built around independent negotiation or a benchmark challenge, not just subscription tools. A campaign could show how to test whether a benchmark's savings target is too low for a specific deal.

This simple framework helps newcomers assess where a category sits before choosing an intervention.

Mark Coulson lays out five stages for any procurement spend category.

• At stage one, nobody owns it; spend simply happens.
• At stage two, someone owns it but only reacts to problems.
• At stage three, there is a plan, reviewed at most once a year.
• At stage four, the category is actively managed against the market.
• At stage five, it becomes a genuine source of advantage.
• Most categories in most businesses sit at stage two: owned on paper, invisible in practice.
• Moving from stage two to stage three does not require more headcount; it takes someone with the time and expertise to run the category.

For brands planning a campaign

A B2B SaaS company reviewing its own procurement could use this as a self-diagnostic before a campaign. The framework helps identify a reactive, unmanaged category and propose a category management program or interim expertise to move it towards active market management.

The post names the risk of accepting a first answer and offers three questions to test judgment.

Mark Coulson recounts how a procurement director's team member pasted a question into Copilot and sent the answer straight to the director's inbox without reading it.

• The director is scared of the technology but also sees enormous potential, Coulson says.
• The post argues that AI does not deskill teams by itself; accepting its first answer does.
• To test the person handing over an AI answer, Coulson suggests three questions: what did it get wrong, what would you change before a supplier saw it, and what did you know that it did not?
• The manager who forwards the answer is a bigger problem than the person who pasted it.

For brands planning a campaign

For a B2B SaaS company buying software or services, this points to a content or training angle on AI-assisted procurement governance: drafting can be fast, but supplier-specific judgment still matters. It could support a campaign about keeping experienced procurement oversight in the loop rather than handing decisions to a tool.

The post makes a common board-level visibility problem easy for newcomers to picture.

Mark Coulson describes watching a CFO review a board pack, scrolling for the procurement line.

• There is none.
• Procurement is folded into other operating costs and spread across a dozen categories, invisible until something moves enough to prompt a question.
• Marketing has a line, IT has a line, but procurement is a rounding error spread across everyone else's budget: present everywhere and owned nowhere.
• Coulson calls it a visibility problem, not a status problem.
• You cannot defend a number nobody can see or get credit for saving one.
• Functions taken seriously in that room make their numbers visible on purpose.

For brands planning a campaign

A campaign for procurement leaders could focus on making spend and savings visible in board-level reporting. For a B2B SaaS buyer, the post also suggests asking how procurement performance is measured and whether supplier spend is genuinely visible before assuming current management is working.

The post turns a common contract term into six real answers and a clear test.

Mark Coulson says six media agencies were asked one question: will you give back every rebate, bonus and discount earned on the client's money?

• Media owners reward agencies for the volume they place, and that reward comes from the client's budget.
• If it is not handed back, part of the approved spend funds someone else's margin and shapes where ads run.
• The six answers fell into three groups: three said yes in writing; two said yes, but returned visible rebates while margin reappeared in discretionary value arrangements; one incumbent declined five transparency terms outright.
• Coulson advises asking in writing before the pitch and making it a clause marketing and procurement own together.

For brands planning a campaign

For a B2B SaaS company procuring marketing or media services, this post supports a procurement-led contract review focused on rebate transparency and total cost, not just agency fees. A campaign could offer a written transparency question or clause checklist to use before any pitch.

After reviewing its data on BMC software, the tool told the client another 15% was achievable.

• Coulson handed the negotiation to an IT procurement specialist with 20+ years of experience, who ignored the benchmark and delivered a 44% saving, worth more than £0.5m beyond what the tool projected.
• They do not dismiss benchmarking outright: the argument is that benchmark platforms cannot weigh the factors that determine how far a supplier will actually move, making their licence fees questionable value alongside an experienced negotiator.

A client story with two comparable savings figures, followed by an open question inviting other practitioners to share their experience.

For brands planning a campaign

For a B2B SaaS company trying to cut software and services costs, treat benchmark output as a starting point, not an outcome, and keep a person accountable for the final number. Coulson's framing gives vendors and consultants an opening to sell negotiation work alongside benchmarking tools.

Examples: Source 1

Most categories in most businesses, they observe, sit at stage two: owned on paper and invisible in practice. In their account, moving from stage two to three requires not more headcount but someone with the time and expertise to run the category, not just hold the title.

A numbered maturity model followed by a self-assessment question about the reader's largest category.

For brands planning a campaign

This gives buyers a diagnostic to run before a procurement engagement and anyone pitching category management work a concrete way to show where the gains lie. A brief built around "which stage is our software category at" is more decision-ready than one built around cost reduction in general.

Examples: Source 1

Marketing has a line, IT has a line, and procurement is folded into "other operating costs" across a dozen categories, visible only when something moves enough to prompt a question.

• They argue this is a visibility problem, not a status problem: you cannot defend or take credit for a number nobody can see.
• Functions taken seriously in that room, they write, make their number visible on purpose, before anyone asks where it is.

A firsthand scene from a board-pack review, leading to a closing question about where procurement's number sits in the reader's own board pack.

For brands planning a campaign

For anyone selling to procurement leaders, the pitch is not better procurement but a procurement number the CFO can see. The same invisibility lets fragmented software spend go unmanaged at SaaS companies.

Examples: Source 1

The director is both frightened of AI and convinced of its potential.

• Coulson's line is that AI is not deskilling the team; accepting its first answer is.
• The cost, they argue, is not the occasional wrong output but the erosion of the judgment people were hired for.
• They offer three questions to put to anyone handing over an AI answer: what did it get wrong, what would you change before a supplier saw it, and what did you know that it didn't.
• Copilot can draft a clause in seconds, but it does not know which supplier slips at year end or which stakeholder will block a change.

A workplace anecdote, a short aphorism, three reusable questions and a direct reader question about disclosure.

For brands planning a campaign

For a SaaS company buying or evaluating procurement technology, the expectation is that tools assist rather than replace the buyer's own context. A vendor whose product depends on customers supplying supplier history and stakeholder knowledge fits this argument better than one claiming to automate the decision.

Examples: Source 1

The answers split three ways.

• Three said yes in writing, two said yes but kept margin inside "value" arrangements priced at the agency's discretion, and the incumbent declined five transparency terms outright.
• Their reasoning is that media owners reward agencies for volume placed, so an unreturned rebate means part of the approved budget funds someone else's margin and steers where ads run.
• A low agency fee is rarely a bargain; the cost reappears where the client cannot see it.

A structured comparison of six real answers grouped into three response types, with reusable group definitions and a closing yes/no/not sure question.

Coulson's posts concern media buying, not software licensing. Ask for software-specific evidence before applying the principle to software procurement.

For brands planning a campaign

For a company procuring marketing and media services, put the rebate clause in writing before the pitch, not after. The same transparency logic—asking where the supplier's margin actually comes from—applies to software resellers and managed-service contracts.

Examples: Source 1

Their team ignored the benchmark’s 15% savings estimate and negotiated a 44% saving. The post identifies BMC as the vendor in the negotiation, not as a sponsor, client or partner.

View example

His posts argue that a guarantee can be an average calculated on a denominator the supplier defined, and that spread pricing, rebate reclassification and specialty margins act as invisible taxes on earnings.

• Coulson’s benchmark story makes the same point in a different category: the tool’s promised saving was real only on the supplier’s terms, and the real number came from someone who read the deal and pushed.
• Both writers treat the contract as the object of analysis and frame unread terms as a governance failure rather than an accident.

Shrack stays inside pharmacy benefit management, writing about modelled exposure across named contract provisions and using quantified, deck-style theses aimed at benefits leaders and CFOs. Coulson ranges across software licensing, category management and media buying, teaching through a client story or brief comparison that ends with a question for the reader.

Read original · Original 2

He shows a merchant who believed they paid 2.9% + 30c but was really above 4.3% once cross-border, FX and gateway fees were counted.

• He asks what a real effective rate means when software, not a person, audits every basis point.
• Coulson makes the equivalent case for agency rebates—a low fee is rarely a bargain because the cost reappears where the client cannot see it—and for a benchmark used as a ceiling.
• Both give readers a diagnostic they can run on their own figures.

Gerada writes as a builder of payment infrastructure, promoting his own platform and pricing it in the post.

• Coulson writes from the buyer’s side as a procurement practitioner, with no product to sell.
• They present the fix as an experienced person plus a written clause, not a new rail.

Read original · Original 2

He argues that AI purchasing behaves like buying raw materials, with token spend moving daily, pricing changing every few months and commits chewed through in a quarter.

• Procurement, he says, should be the team finance and engineering call first when costs move.
• He also dissects the moment a vendor stops moving and shows how to distinguish a genuine limit from a stalled negotiator, using a buyer who found a further 21% saving after the rep said there was no more room.
• Coulson’s position is that procurement earns its place through exactly that supplier knowledge, which no drafting tool holds.

Etkin writes from inside a procurement-intelligence vendor, orienting the argument toward a platform and its data for SaaS and AI spend.

• Coulson writes independently and more broadly, covering category maturity, board reporting and media rebates.
• Their recommendation is usually a question for the reader to put to their own supplier, rather than a tool to buy.

Read original · Original 2

The Carta account is the posts' clearest example: the vendor had stopped moving, Carta's team learned through Tropic's intelligence that the vendor was stalling rather than out of room, changed approach and found 21% in additional savings.

• The posts also make a case about AI consumption, where token spend moves daily, pricing changes every few months, and the model a buyer signed for can be deprecated before the contract ends.
• For a B2B SaaS company whose bills now include AI usage, that is the part of the estate where the existing renewal calendar and category owner no longer fit.
• If the campaign needs one negotiation story, a number on the table, and a route into spend that procurement does not yet own, these posts carry it more directly than a general category argument.

The negotiation evidence sits inside a vendor's marketing.

• The Carta saving is attributed to Tropic's intelligence product, the post closes with a Tropic case-study link, and the AI material promotes a named webinar.
• The outcomes are therefore product claims from an interested party, not independent accounts.
• A campaign built on this material is closer to a vendor case study than a neutral view of whether to purchase negotiation intelligence.
• If a client needs an independent assessment, ask for practitioner evidence beyond these posts.

Examples: Source 1 · Source 2

The posts model an agreement across 15 provisions, put figures on the leakage and specialty-margin exposure, and argue that the strongest negotiation starts before the RFP with provision-level details and replacement wording.

• Audit rights, rebate definitions and guarantee mechanics are treated as the product being bought.
• For a software or services deal with a complex schedule of fees, rebates, exclusions and true-up clauses, that method—read the clause, quantify it, draft the replacement—takes the job Coulson describes in their benchmark post down to the clause level.
• Shrack also sells a fixed-price contract review as a fast deliverable, which suits a campaign that has to give the reader something concrete to run this week.

The subject distance is the problem.

• Every post is about a pharmacy benefit manager contract in US healthcare: spread pricing, rebate reclassification, average wholesale price, BER guarantees.
• Nothing in the posts covers software licensing, SaaS renewals or services category management, so a B2B SaaS buyer can borrow the method but not the examples.
• Presenting PBM leakage as software-procurement insight would make a claim the posts do not.
• The posts also sell a paid contract-clarity report, weakening the independent-buyer framing of Coulson's benchmark-versus-experience argument.
• If audience data matters to the campaign, ask Shrack for it.

Examples: Source 1 · Source 2

One post takes a single merchant, shows a quoted 2.9% plus 30 cents becoming a blended rate above 4.3% once cross-border assessments, FX spread, gateway fees, chargeback fees and downgraded interchange are added, and converts the gap into $1.54M a year.

• It names the published fee schedules behind the gap and gives the reader a five-minute diagnostic: total fees divided by total processed volume.
• That is the same close reading Coulson applies to a benchmark number, aimed at a different cost line.
• The second post argues that agentic checkout changes who reads the fee schedule, and why a per-transaction toll that is a rounding error at human volume becomes a real tax that gets engineered around.
• Choose Gerada when the campaign is about fee legibility and machine-driven purchasing, not supplier category management.

Scope narrows to one category.

• Payments is not software or services procurement, and these posts do not cover vendor selection, category stages, supplier relationships or board reporting.
• The commercial copy points to a specific payment product with its own rate stated in the post, while the quantified merchant example comes from the company selling the alternative.
• The diagnosis and remedy therefore come from the same side of the table.
• For the buyer's stated need, Gerada means rebuilding the campaign around payments rather than answering the original question.
• Treat the product claims as commercial copy, not independent findings.

Examples: Source 1 · Source 2