Mid-Market ERP Selection Advisory, Run on Arithmetic Instead of Demos
A fractional CIO who sells no software sits on your side of the table, sizes the decision with numbers your own systems already produce, and hands you a shortlist you can defend to a board.
A mid-market ERP selection advisory puts an independent fractional CIO in charge of your evaluation, sizing requirements, shortlist and honest year-one cost from your own transaction data before any vendor books a demo.
Last updated: August 21, 2026
Mid-market ERP selections rarely fail on the software. They fail on a number nobody added up.
The shape is familiar enough that you can set your watch by it. Four finalists. Three rounds of demos. A scoring sheet long enough to need its own tab, and a steering committee that meets every other Thursday for most of a year. At the end of it every finalist lands within a few points of every other finalist, because the sixty rows on that sheet are things all four of them do.
So the decision gets made on whichever sales team was warmest.
Then year two arrives. The volume the business actually runs turns out to sit two license tiers above what got quoted. An integration call limit nobody read starts throttling the nightly sync. The number in the board deck stops matching the number in the system, and a finance team that was promised one source of truth is back in a spreadsheet by the second close.
None of that was unknowable. All of it was countable in week one.
KORE1 has placed IT and finance systems talent since 2005, and this advisory sits inside our wider ERP consultant staffing practice. It covers one thing. The decision itself, before you’ve signed anything.
What a Selection Advisory Owns, and What It Deliberately Doesn’t
This isn’t a project manager and it isn’t an implementation partner. It’s a small, senior engagement with one job, which is getting you to a decision you can defend, and then getting out of the way. That’s the entire scope.
The work itself is unglamorous. Sit with the people who close the books and the people who ship the product, and write down what the business does in volumes rather than adjectives. Pull the counts. Transactions a month, order lines, SKUs, entities, currencies, integration calls, users who need a full license against users who only need a screen. Turn all of that into requirements a vendor can either demonstrate live or has to concede. No adjectives survive.
Then build the shortlist from the counts instead of from an analyst grid, and write the demo script yourself.
That last piece decides more than the scoring sheet does. Hand a vendor an open hour and they’ll show you the three workflows their product is best at, which is a completely rational thing for them to do and useless to you.
One boundary, up front. This seat takes no referral fee, no reseller margin, and no implementation contract on the far side of the decision. If the honest answer is that you should stay where you are for another eighteen months and fix the process instead, that answer costs us the follow-on work and you get it anyway.
Five Numbers You Already Have, and the Line Each One Has to Clear
The figures below come from a single 2026 Foretopia proposal for a mid-market regulated manufacturer, anonymized and used as a worked example. They are not your numbers. They are the shape of the arithmetic that settles a selection before anybody scores a demo, and every one of them was computable from data the buyer already owned.
Integration calls against the licensed allowance
A unit-level integration design modelled out at roughly 1,350,000 API calls a year against a licensed allowance of 130,000. Ten times over. Batching the same interface at transaction level, about 141 documents a month at four to six calls each, lands under a quarter of the allowance. Same requirement, same platform, tenth of the traffic.
Validation hours against the build estimate
For bespoke software in a regulated environment, validation runs about 30% of base build hours. On a 1,060-hour build that’s roughly 320 hours, and they only appear on the page if somebody asks for them. Somebody rarely asks.
Year-one cash against the software line
Discovery, risk assessment and validation planning came to 182 hours and $44,110 fixed. Build, validation and release landed between $241,577 and $316,735. Then validated-state maintenance across the platform’s twice-yearly releases runs 120 hours and about $31,046 a year, which repeats forever and almost never makes it onto the comparison slide.
Authentication the platform will actually accept
A NetSuite RESTlet won’t take a plain API key. External callers need OAuth 1.0 token-based auth or signed OAuth 2.0 requests. So a vendor whose integration story is that they’ll post to your endpoint with a key has quietly handed you an authenticated relay to build, host and own. That’s a project.
The cheaper proposal against the native one
An integration-platform alternative priced at $265,900 excluding the platform licence, roughly 15% under the native midpoint of $312,530. The gap narrows once you count what stays bespoke inside it, because a realistic build still contains scripted transformation steps that carry the same validation burden wherever they run. The saving shrinks.
Three of those five are settled with a saved search and twenty minutes. Cheap insurance. The other two need somebody who has read a statement of work before, which is most of what you’re buying when you buy this engagement. For the committee side of the same problem, Colin Boothe’s walkthrough of how to choose an ERP for a mid-market business goes deeper than a service page should.
Four Documents That Outlive the Decision
Whatever you choose, and whether or not you choose us for anything after this, these are yours. They stay useful during implementation, and two of them get used again at renewal.
Real volumes by module, entity and interface, with the license tier each one puts you in. The single most useful page in the whole engagement. Start there.
Written so a vendor demonstrates it live or concedes it in writing. Adjectives get struck. The list always ends up shorter than the one you started with.
Your transactions, your awkward edge cases, run in the same order by every finalist so the comparison means something.
Licence, implementation, integration, internal backfill and the recurring maintenance line, modelled across three years rather than one.
What Independent Selection Help Costs
Rates first. They set the shape of everything after. A fractional CIO on an advisory footing runs roughly $175 to $300 an hour in the mid-market, and most selection engagements land between 60 and 160 hours spread across six to twelve weeks. Specialist consultancies bill delivery closer to $215 an hour blended, per Foretopia’s 2026 benchmarks, with a differentiated rate near $275 for an independent validation lead on regulated work.
Now set that against what the decision is worth. Research by McKinsey and the University of Oxford, covering more than 5,400 IT projects, found they ran 45% over budget while delivering 56% less value than predicted, and that every additional year on a project added about 15% to the cost overrun.
Fair warning on that figure. Those were projects with price tags over $15 million, which is well above most mid-market buys. Same mechanism, smaller numbers. The arithmetic on the extra year doesn’t change at all.
“Nobody calls me because the software can’t do it. They call because the thing they bought got sized on a demo, and now the nightly sync is throttled and finance is back in a spreadsheet.”
Duration is where mid-market budgets quietly get eaten. A selection that runs fourteen months costs more in committee hours than the advisory does, and the requirements you’re evaluating against go stale while you deliberate. Twenty-seven years ago in Harvard Business Review, Thomas Davenport warned that an enterprise system imposes its own logic on a company’s strategy. That hasn’t aged a day, and it’s still the question a feature matrix can’t answer.
Where the Hours Go
Six to twelve weeks end to end for a single-entity business, longer if you’re consolidating several. The hour ranges below are typical rather than fixed, and stage one is where the surprises live.
- 01
Intake and Counts
We pull volumes out of your current systems rather than asking what you think they are. Roughly 15 to 30 hours, and the gap between the two is often the story.
- 02
Requirements Worth Testing
Every line rewritten as something a vendor shows you or admits they can’t. About 15 to 35 hours. Mostly deleting.
- 03
Shortlist From the Counts
Two or three finalists, picked because your volumes and entity structure fit them. Four finalists is a committee compromise, not a shortlist.
- 04
Your Script, Their Demo
Same transactions, same edge cases, same order for every finalist, scored the same afternoon while everyone still remembers what they saw.
- 05
The Number and the Recommendation
Three-year totals, the risks we’d accept, the ones we wouldn’t, and a written recommendation you can hand to a board without translating it first.
This Desk Covers the Decision, Not the Build
Once you’ve chosen, the work changes hands and so does the profile you need. We’ll say so. Stretching an advisory into a delivery contract is exactly how the neutrality goes away.
Configuration and delivery belong with NetSuite implementation consultant staffing or the wider ERP consultant staffing bench. If you already have a shortlist and a live RFP in flight, and what you actually need is a contract seat to run that process, that’s ERP vendor selection and RFP advisory staffing. If the whole transformation needs a leader rather than a decision, that’s a fractional CIO for ERP and digital transformation.
Haven’t decided whether to buy at all? The free ERP readiness assessment is a better first hour than a demo, and the ERP comparison tool will orient you on NetSuite, SAP, Dynamics 365 and Epicor without a sales call. If you’re leaving QuickBooks specifically, that move has its own shape and a different set of counts.
Reviewed by Colin Boothe, CIO at Foretopia
Colin has spent eight years in ERP and business operations consulting, most of it inside NetSuite, connecting ERP, WMS and AI stacks for companies that grew faster than their systems did. He stood up a full tech stack for a wholesale and ecommerce business scaling past $250M in about seven months. Authority from shipping, not credentials. He describes the fractional model as a rentable CIO, on the theory that the companies who need his judgement most can rarely justify it full time.
The delivery benchmarks, the API allowance arithmetic and the contract-scope detail on this page come from his practice, anonymized. The placement figures and rate ranges are ours. His standing position is that the software is almost never the hard part, that the org chart and the process usually are, and that most companies are running a fraction of what they already paid for.
Three Ways Companies Buy This
The advisor and the method don’t change. What moves is how long you keep the seat and who signs the recommendation. That’s it.
Scoped Advisory
A fixed-scope selection engagement priced to the window, typically 60 to 160 hours across six to twelve weeks, ending in a written recommendation.
Project Staffing →Fractional and Interim
A part-time CIO on a KORE1 W-2 who runs selection first, then stays through implementation oversight on a set number of days a month.
Contract Staffing →Hire the Seat
Companies buying an ERP they’ll run for a decade often decide mid-engagement that the role should be permanent. We’ll search for it.
Direct Hire details →Common Questions
How do you choose an ERP for a mid-market business without a nine-month bake-off?
Size the business in real volumes first, turn those counts into requirements a vendor must demonstrate live, then shortlist to two or three finalists and run every one of them through a demo script you wrote.
The bake-off gets long when the shortlist is built before the counts are. Four finalists on a sixty-row feature matrix will always score within a few points of each other, so the process keeps running until somebody gets tired. Cut to two finalists chosen on volume fit and the whole thing compresses. Weeks, not quarters.
What does a mid-market ERP selection advisory cost?
Most selection engagements run 60 to 160 hours over six to twelve weeks, at roughly $175 to $300 an hour for a fractional CIO, so a typical scoped advisory lands somewhere between $15,000 and $45,000.
Compare that against a single recurring line item. On one 2026 proposal we’ve worked from, validated-state maintenance alone came to about $31,046 a year, every year, and it wasn’t on anybody’s comparison slide. Finding one line like that pays for the engagement in month one.
Isn’t this what our implementation partner does anyway?
Not the selection part, and not neutrally. An implementation partner is compensated on the far side of the decision, so their advice about which platform to buy is shaped by which platform they’re certified to deliver.
That’s not a knock on them. It’s the structure. They’re good at the build and they should be doing the build. Asking a reseller which system to buy is like asking a surgeon whether you need surgery, and the honest ones will tell you the same thing.
We already have a shortlist and demos booked. Too late?
No. It’s the most common entry point we get. A compressed version runs the counts and rewrites the demo script in about two weeks, which is usually enough to change what you ask the finalists.
Roughly a third of the time the counts also change the shortlist, and once in a while they eliminate a finalist entirely on a licensing tier nobody had checked. If your demos are next week, we’d still rather look at the volumes first.
Will we just be told to buy NetSuite?
No. The advisor takes no reseller margin and no implementation contract afterward, so there is no commercial reason to steer you, and the counts decide the shortlist rather than the advisor’s résumé.
Deep platform experience does matter, and it cuts the other way from what people expect. Somebody who has implemented a platform knows precisely where it’s weak, which is exactly the knowledge a vendor demo is designed to route around. Plenty of mid-market companies end up on Dynamics 365, Acumatica or Sage Intacct, and some are better off staying put and fixing process for a year.
How long should ERP selection actually take?
Six to twelve weeks for a single-entity mid-market business, and twelve to twenty for a multi-entity group with consolidation and intercompany requirements. Anything past six months is usually a decision-rights problem rather than an analysis problem. Name who decides.
Length is expensive in a way that doesn’t show up on an invoice. McKinsey’s numbers put roughly 15% of extra cost overrun on each additional year a project runs, and your requirements drift while the committee deliberates. The spec you started with in January describes a business that no longer exists in November.
What happens if the recommendation is that we shouldn’t buy anything yet?
You get that recommendation in writing, with the conditions that would change it. It happens on a meaningful share of engagements, usually when the process problem underneath the software problem hasn’t been solved.
A new ERP will happily automate a broken close. Faster, too. In every direction, with much better reporting on how broken it is. If the counts say your data quality or your month-end process is the constraint, spending eighteen months fixing that and then buying is the cheaper sequence, and the free ERP readiness assessment is where that conversation usually starts.
Send us your transaction volumes and your entity list. We’ll tell you on one call whether you need an advisory or just a better demo script.
No pitch deck, no reseller margin, and a straight answer if the honest recommendation is to wait.
Scope a Selection Advisory →