Choosing the systems integrator is a separate decision from choosing the software, and it carries more of the risk. The SI is 40 to 60% of what an ERP program costs to implement, per ERP Research, an independent research firm, and it is the largest single driver of whether the program goes live on the date you were promised. Choose on four things: the named team you will actually get, references from programs of your size, industry, and scope, a rate card you can read, and a contract that holds the SI to both. The logo on the proposal predicts almost none of this.
Three kinds of SI, and the right one depends on your size and your vendor
Three kinds of firm bid ERP work: a global firm, a boutique specialist, and the vendor's own professional services arm. Each fits a different shape of program, and the mistake is treating them as interchangeable bidders on one price. The companion guide to selecting a systems integrator walks the full shortlist-to-signature sequence; this article goes deep on the four tests that decide it.
| Kind of SI | Strength | Risk it carries | When it is the right answer |
|---|---|---|---|
| Global firm | Bench depth across countries and modules, able to staff every region | Highest blended rate, and the team that wins the bid is often not the team that delivers | Multi-country tier 1 rollouts (SAP S/4HANA, Oracle Fusion) at 10,000 or more employees that need staff in every time zone |
| Boutique specialist | Deep knowledge of one product, named senior consultants, a lower blended rate | Thin bench if a key person leaves, limited reach across geographies | Single-region tier 2 programs (NetSuite, Business Central, Epicor Kinetic) where product depth matters more than global coverage |
| Vendor professional services | Closest to the product and its roadmap, useful when few partners are trained yet | Less industry and process depth, and the least independent voice when the software itself is the problem | A clean cloud rollout on a product with a thin partner network, or the first wave of a newly released module |
Large programs often mix the three. A global firm runs as prime contractor while a boutique owns a specialist module and the vendor's own people handle a new component its partners have not learned. The rule we recommend is to match the firm to the hardest part of your program, not to the whole, and to price the mix rather than defaulting to one logo for everything.
Evaluate the named team you will get, not the logo on the proposal
A proposal is won by the firm's best people and delivered by its most available ones. The brand name tells you the firm can field a good team somewhere. It does not tell you that team is coming to you. The gap between the two is where most disappointment lives.
Require named key personnel for the lead roles: the program lead, the lead solution architect, the data lead, and the integration lead. Ask for their CVs, ask what else each is staffed on, and ask for the percentage of their time your program holds. Then require your written approval to replace any of them. The common failure without that clause is that the named experts win the work and roll off within the first weeks after go-live, replaced by juniors you are still billed for at the blended rate. A program that drifts far enough to need a rescue costs 30 to 50% of the original budget, per Deliberate Directions, a firm that sells NetSuite rescues, so treat it as a single firm's planning range. A quiet team swap is one of the ways a program drifts there.
In our experience a tier 1 program needs 36 months from decision to go-live. The lead architect you were pitched should be contracted to stay through the build that spans most of those months, not just the pitch that fills the first few weeks.
References from programs of your size, industry, and scope are the strongest predictor
If one test decides the choice, it is this one. Gartner predicts that by 2027 more than 70% of recently implemented ERP initiatives will fail to fully meet their original business case goals. The SIs that stay out of that number are usually the ones that have run a program of your shape before. A reference that matches on the vendor alone is close to worthless, because the same software behaves nothing alike across sizes and industries.
Three dimensions have to match for a reference to count. Size means revenue, user count, and legal entities, because a cost added per legal entity runs 10 to 20% of implementation and a firm that has only done single-entity work has never carried that weight. Industry means your regulatory and process shape, not a loose sector label. Scope means the same modules, a similar count of integrations, and the same global-versus-single-region reach.
| Dimension | A reference that tells you little | A reference that counts |
|---|---|---|
| Size | A $300M single-entity client when you are $3B with 20 entities | Within roughly the same revenue band, user count, and number of legal entities |
| Industry | A retailer when you are a regulated utility | Your industry, with the same compliance and process demands |
| Scope | Finance-only when you are rolling out finance, supply chain, and manufacturing | The same modules, a similar integration count, and the same geographic reach |
Call the reference and ask to speak to its business process owner, not only the CIO, because the owner lived the testing and the change orders. Ask whether the team that delivered that reference is the team you were pitched. Ask what went wrong, because every program has a bad month, and ask how the SI priced and handled the first real change order. A reference that cannot describe a problem is a reference that has been coached.
Rate transparency and the junior-senior mix decide what you actually pay
A blended rate is one average across all roles. A fully loaded rate names each role and its cost. The gap between them is the junior-senior mix, and that mix is the single largest rate lever the SI controls. A proposal that shows only the blend has hidden the lever that moves its own margin.
KORE1, a staffing firm, puts 2026 boutique SI rates at $150 to $250 an hour for Dynamics 365, $150 to $275 or more for NetSuite, and $200 to $325 or more for SAP, with one of its own pages showing less senior roles at $90 to $175. These are confidence C figures from a firm that places consultants, and KORE1's pages disagree with each other, so treat them as directional ranges rather than fixed prices. The spread inside each product is the point: the top of the range is roughly 1.5 to 1.7 times the bottom, and that gap is the junior-to-senior split you are paying for.
Consider a $4M fixed fee built on 20,000 hours at a blended $200. That blend could be 30% senior at $300 and 70% junior at $130, which costs the SI $181 an hour to deliver. The $19 gap across 20,000 hours is $380,000 of margin the blended rate never shows. Margin is not the problem. Not being able to see it is. Ask for the rate by role, ask for the junior-senior split, and ask which of the named people from the last test sit in the senior rows. Design and fit-gap work should run 10 to 20% of total hours, per ERP Research; a design line below that means the SI plans to work out your requirements while billing you to build.
Fixed-fee, time-and-materials, and the hybrid that holds, with AI claims in the hours
No single pricing model fits a whole program. Fixed-fee protects you where scope is firm and punishes you where it is not, because the SI either prices the uncertainty into the fee or claws it back through change orders. Time-and-materials protects the SI and exposes you to open-ended hours. The hybrid that holds up draws the line at scope certainty.
| Model | Fits | Risk it creates | Where to use it |
|---|---|---|---|
| Fixed-fee | Firm, well-understood scope | Padding, or change orders when scope shifts | Standard configuration, training, a defined cutover |
| Time-and-materials | Genuinely uncertain work | Open-ended hours, weak cost control | Data cleanup, discovery, integrations to unknown systems |
| Hybrid | Most tier 1 and tier 2 programs | A blurred line between the two | Fixed on the firm parts, T&M with a cap on the rest |
The rule we recommend is fixed-fee on anything you can specify today, T&M with a not-to-exceed cap on anything you cannot, and the line between them written into the contract rather than left to the SI. How to pull the hours apart inside either model is the work of reading the SI proposal line by line.
AI efficiency claims belong in whichever model you pick, and they belong in the hours. SIs now say agents cut their effort, and the claims are large: McKinsey, writing in May 2026, projects at least a 50% cut in implementation effort, while BCG in April 2025 put it lower at 20 to 40%. A saving that is real reduces the hours in a work package. A saving that is marketing stays a percentage on a slide. The rule we recommend is to treat any AI claim with no hours delta by work package as worth zero in your price comparison, and to put a re-baseline point at the end of design where the SI shows actual hours against what it quoted.