Through the integrator

Most industrial technology reaches the plant through someone other than the company that built it. System integrators (SIs) and OT integrators, automation distributors, value-added resellers (VARs), machine builders, engineering firms, managed service providers and cloud marketplaces all sit between a technology vendor and the plant. Across IT as a whole, Canalys forecast in November 2024 that partner-delivered IT would account for about 70% of the total addressable IT market in 2025, and Omdia, of which Canalys is now part, has since put the partner share for 2026 at about two-thirds of IT spending [1]. In our experience, industrial buyers depend on partners at least as much, because someone has to connect a new system to equipment and controls that are often decades old.

A vendor's partner plan usually rests on a short list of assumptions: the partner holds the relationship with the plant, the partner can implement the product, the partner earns enough to prioritize it, the partner route will not fight the vendor's direct sales or the partner's other suppliers, and the plant will accept the partner as accountable after go-live.

These assumptions are rarely tested before the vendor commits headcount, incentives and a launch date.

This paper sets out how to test each one with partners and with end buyers, what the evidence looks like at each grade, and how to decide between a distributor-led, SI-led, OEM-embedded, co-sell, marketplace, developer-led or direct route. It is written for go-to-market and alliance leaders at technology vendors, and for OEMs building a software or service business. Who approves the purchase inside the plant is covered in The committee nobody mapped. This paper covers the route to that committee.

Three features of industrial buying make partners hard to avoid.

Integration. New software or hardware has to connect to existing controllers, networks and plant systems, often from several suppliers and several decades. In our experience, the people who can do that work at a given plant are frequently the integrator who installed the existing systems.

Trust. Plant leaders are cautious about who touches production systems. An integrator with years of work at a site has a trust the vendor does not, and, in our experience, plant leaders often ask their integrator's opinion before they talk to a new vendor.

Coverage. A vendor cannot staff sales and service teams near every plant it wants to reach. Distributors, integrators and machine builders already have people in those regions.

The partner types differ in what they offer:

Partner type	What they bring	What they usually want
System integrator or OT integrator	Plant relationships, integration skills, project delivery, usually within one region	Project and services revenue, products that make projects easier to win
Automation distributor or VAR	Coverage across territories, often exclusive by brand, existing accounts, logistics	Margin on product, rebates tied to volume, repeat sales, low support burden
Machine builder or OEM	The machine the software runs on, the installed base	Differentiation for the machine, service revenue, control of the customer relationship
Engineering or construction firm	Early influence on new plants and major upgrades	Specifications that reduce project risk
Managed service provider	Ongoing operations after go-live	Recurring revenue, standard products
Cloud or software marketplace	Procurement convenience, use of committed cloud spend (cloud-deployed offers only; on-premises licenses may not count, depending on the provider's eligibility rules [9] [10])	Transaction volume
Developer or integrator self-serve	A free download tried before anyone sells; adoption evidence	Someone in the plant who can turn usage into a budget

A partner plan that treats these as one "channel" misses the fact that each one wins in a different way.

Most partner plans rest on five claims. Each can be tested before the vendor commits.

1. The partner holds the relationship. The plan assumes the partner can open a door to the people who sign at the plant. The question is whether the partner knows the budget holder or only the maintenance and engineering staff who call when something breaks. Those are different doors.

2. The partner can implement. The plan assumes the partner has, or will build, the skills to deploy the product. Building a practice costs the partner time and money: training, certification, a first project done at a loss. The question is whether the partner will make that investment for this product.

3. The partner earns enough. The plan assumes the product is worth the partner's attention. A partner compares every product with what it already sells. If the new product earns less per hour of effort, or takes sales time from a product that earns more, the partner will sign the agreement and not sell.

4. The route does not fight itself. The plan assumes the partner route will coexist with the vendor's direct sales, with other partners, and with the partner's existing suppliers. Where a direct salesperson and a partner pursue the same account, or where the product competes with a line the partner already carries, conflict follows.

5. The plant accepts the partner as accountable. The plan assumes that when something goes wrong after go-live, the plant will call the partner and accept the partner's answer. Some plants will. Others will insist on the vendor, which changes the support model and the cost.

A signed partner agreement shows that the first conversation went well. It does not evidence any of the five.

The relationship is tested from the plant's side, not the partner's. Interview a sample of the plants the partner claims as accounts. Ask who at the plant they talk to, how often, and whether the partner has ever brought them a new product that they bought. A partner who is trusted for maintenance may have no influence over a new capital or operating budget.

• Verified: Budget holders at named plants confirm that the partner has brought them new products they bought, and would take a meeting on this one.
• Benchmarked: The partner's account list and its own description of its relationships.
• Assumption: The assumption that a partner with many accounts has influence at each.

The capability is tested with the partner's delivery team, not its leadership. Ask the engineers who would deploy the product what they know about it, what they would need to learn, and how long the first deployment would take. Ask partner leadership how many people they would train, and when.

• Verified: Named delivery staff, a training plan with dates, and a first project the partner will do at its own risk or on agreed terms.
• Benchmarked: A partner commitment to "build a practice" without names or dates.
• Assumption: The assumption that a partner with related skills can deploy the product.

Partner economics decide whether the partner sells. The vendor's view of the economics, usually a margin percentage, is often not the partner's view.

A partner thinks about three things:

• Revenue per hour of effort. How much the partner earns from a deal, including services, compared with the sales and delivery hours it takes.
• What the product replaces. If it displaces a product the partner already sells, the partner loses that margin. If it creates new services work, the partner gains.
• Time to first profit. A new product usually costs the partner money on the first projects. The question is how many projects until it earns a return.

The strongest evidence comes from asking partner sales and delivery leaders to estimate these for a realistic first deal, and comparing the estimate with a deal they closed recently for another product. If the new product earns less per hour than the partner's existing work, the plan needs a different incentive or a different partner.

For OEMs moving to service or outcome models, the channel's economics can change more sharply. A dealer's margin on equipment and parts may disappear under a direct service model. The servitization bet covers this in more detail.

Channel conflict is predictable. The question is where it will happen and how often.

Research on hybrid channels, in which a supplier sells both directly and through partners, found that the frequency of conflict, not its intensity, reduced channel performance [2]. In practice that means small, repeated disputes over accounts, pricing and credit do more damage than a single large one.

Common sources of conflict in industrial partner routes:

• Direct sales and partners in the same accounts. Especially large accounts, where the vendor wants a direct relationship and the partner has the history.
• Several partners in one region. Two integrators bidding the same plant with the same product at different prices.
• The partner's existing suppliers. An automation distributor that carries a major controls brand may be restricted, formally or informally, from promoting a product that competes with that brand's software.
• Marketplace and partner sales of the same product. A customer can buy through a cloud marketplace and bypass the partner who did the pre-sales work.
• Distributor territory against integrator region. An exclusive distributor territory rarely matches the region an integrator covers.

The test is to write down, before launch, the rules for each case: which accounts are direct, how partners register deals, how credit is shared, and what happens when the rules conflict. Then ask partners whether they would accept those rules. Their answer is evidence. So is the question they ask first.

The plant's view of accountability decides the support model.

Ask plant leaders directly: when this system fails at two in the morning, who do you expect to call? If the answer is the partner, ask whether they would accept the partner's diagnosis and repair. If the answer is the vendor, the vendor needs a support organization that can reach the plant, regardless of who sold the product.

In our experience, many plants give a mixed answer: the partner for the first call, the vendor for anything the partner cannot fix. That is workable if the handoff is defined. It fails when each side assumes the other is responsible, and the plant finds out during an outage.

The answer also affects price. A plant that insists on vendor accountability will expect the vendor's support to be included, and the partner's margin has to fit around it.

Once the five claims have been tested, the evidence points to a route; a marketplace listing sits alongside any of them.

SI-led. The partner owns the relationship, the sale and the delivery. The vendor supports the partner. This works when the partner holds the plant relationship, has or will build the capability, and earns enough. It scales coverage quickly and gives the vendor less control over the customer experience.

Co-sell. The vendor and the partner pursue accounts together: the vendor brings product knowledge and executive access, the partner brings the relationship and the delivery. This works when the product is new or complex and the partner's capability is still developing. It costs more per deal and needs clear rules on credit.

Direct, with partners for delivery. The vendor sells and owns the relationship, and partners deliver the integration work as subcontractors. This works when the vendor needs control of the customer relationship, for example because the product is strategic or the pricing is complex, and partners want services revenue more than product margin.

Distributor-led, SI-delivered. The distributor transacts and holds the account across its territories; regional integrators deploy. This works when the product sells alongside the brands the distributor already stocks and the integrators in each region will take the deployment work. The territory and the region rarely match, so the plan has to say who owns the account where they overlap.

OEM-embedded. The vendor's product is built into an equipment maker's offer, and the machine carries it to the plant. This works when the OEM's product team designs it in, a billing owner exists, the vendor is paid per unit or per subscription attached, and the OEM's dealers or service arm will carry it. The OEM owns the customer and the renewal, so the plan has to say what the vendor learns about usage and when.

Marketplace. A hyperscaler or industry marketplace transacts, and the listing sits alongside any of the routes above. This works when the buyer's committed cloud spend can absorb the purchase, which may not count on-premises licenses, depending on the provider's eligibility rules [9] [10], and the co-sell rep's quota rewards it. It bypasses the partner who did the pre-sales work unless the plan pays that partner anyway.

Developer-led. A free download reaches an engineer in the plant or at an integrator before anyone sells. Downloads are easy to count and prove little. The test is whether the usage reaches someone who can turn it into a budget, and how often a trial becomes a purchase order.

What IT's channel did

The IT channel has already been through a version of the shift an industrial vendor is planning. Omdia forecasts enterprise software sales through hyperscaler cloud marketplaces growing from $30 billion in 2024 to $163 billion by 2030, with partners facilitating nearly 60% of those transactions by 2030 [7]. From interviews with managed service providers, technology services distributors and advisors, Omdia also reports that co-selling has matured from sporadic gap-filling into planned go-to-market motions [8]. Industrial vendors have started down the same path, but in our experience the industrial channel is still paid, for the most part, per transaction and per project.

Four tests for a recurring offer

Where the product is priced as a subscription, by usage or by outcome, four more questions decide whether the channel will carry it:

• Will the distributor sell, not only quote, a subscription?
• Does the integrator have the capacity, in each region the rollout needs, to deploy and support it?
• Who holds the renewal: the vendor, the distributor or the integrator?
• Do the rebates and market development funds that move the distributor's sales team survive a recurring price?

For a platform company entering an industrial vertical, the question often becomes whether to build a direct vertical team or rely on partners who already know the industry. A partner route reaches more plants sooner. A direct team learns more about the buyer and keeps the relationship. The evidence from the five tests usually decides it: if partners do not hold the budget relationship, or will not invest in the capability, an SI-led route will not work at the speed the plan assumes, however many agreements are signed. The same evidence tells a planning review whether to fund partner managers or a direct vertical team, and in which regions.

A partner route is tested from both sides.

End buyers first. Budget holders (the plant manager or VP of operations), the plant controller who checks the value case, the maintenance and reliability manager, and OT security at a sample of target plants. They answer which partners they trust, for what, and who they expect to be accountable. Starting here keeps the partner interviews honest.

Partner leadership. On strategy, economics, the practice they would build and the rules they would accept.

Partner sales and delivery staff. On what they would sell, what they know about the product, and how long the first deployment would take. Their answers often differ from leadership's.

The vendor's own sales team. On which accounts they expect to keep direct, and how they are paid when a partner closes a deal.

Each interview is scored on the same dimensions, and each of the five claims receives a grade of Verified, Benchmarked or Assumption. The evidence standard is set out on the method page, and what we test for technology companies covers vendors selling into industry.

End buyers

1. Which partners do you work with today, and for what kind of work?
2. Has a partner brought you a new product that you bought? Who approved it?
3. Would you take a meeting on this product if your partner brought it?
4. When the system fails at night, who do you expect to call?
5. Would you accept the partner's diagnosis, or would you want the vendor involved?

Partner leadership

6. How many of your customers have asked for this kind of product?
7. How many people would you train, by when, and who pays for the first project?
8. What does this product earn you per hour of effort compared with what you sell today?
9. Does it compete with any product you already carry, or with any supplier's software?
10. Would you accept these rules for account ownership, deal registration and credit?

Partner sales and delivery staff

11. What do you know about this product today?
12. How long would the first deployment take, and what would you need to learn?
13. Which of your customers would you take it to first, and why?

The vendor's own team

14. Which accounts will stay direct, and who decides?
15. How is the direct sales team paid when a partner closes a deal in its territory?

1. Canalys (now part of Omdia), November 2024 IT Opportunity forecast: worldwide IT spending of US$5.44 trillion in 2025, with partner-delivered IT at 70% of the total addressable market. https://omdia.tech.informa.com/om142315/it-spending-to-expand-8-in-2025-partnerdelivered-it-to-account-for-70 Later Omdia figures put the 2026 share at about two-thirds: 66.7% in 2026, down from 70.1% in 2025, in its January 2026 update https://omdia.tech.informa.com/blogs/2026/jan/how-will-changing-it-spending-trends-impact-global-channel-chiefs, and 65% as reported by Channel Dive, 21 July 2026. https://www.channeldive.com/news/the-channel-owns-less-than-two-thirds-of-global-it-spend/825825/
2. Kevin L. Webb and John E. Hogan, "Hybrid channel conflict: causes and effects on channel performance," Journal of Business & Industrial Marketing 17, no. 5 (2002): 338 to 356; the abstract states that "the frequency of conflict, but not its intensity, has a negative effect on channel system performance." https://doi.org/10.1108/08858620210439031
3. Rockwell Automation, Form 10-K for the fiscal year ended 30 September 2025. https://www.sec.gov/Archives/edgar/data/1024478/000102447825000116/rok-20250930.htm
4. PTC, Form 10-K for the fiscal year ended 30 September 2024. https://www.sec.gov/Archives/edgar/data/857005/000095017024127231/ptc-20240930.htm
5. Bentley Systems, Form 10-K for the fiscal year ended 31 December 2025. https://www.sec.gov/Archives/edgar/data/1031308/000103130826000007/bsy-20251231.htm
6. Cognex, Form 10-K for the fiscal year ended 31 December 2024. https://www.sec.gov/Archives/edgar/data/851205/000085120525000012/cgnx-20241231.htm
7. Omdia, "Hyperscaler Cloud Marketplace Sales to Hit $163 Billion by 2030," press release, 6 October 2025, read in its syndicated copy: sales "projected to surge from $30 billion in 2024 to $163 billion by 2030" and partners expected to "facilitate nearly 60% of all marketplace transactions" by 2030. https://finance.yahoo.com/news/omdia-hyperscaler-cloud-marketplace-sales-073800236.html
8. Channel Dive, "Co-selling sweeps through the channel," 10 June 2026, reporting Omdia's interviews with managed service providers, technology services distributors and advisors. https://www.channeldive.com/news/co-selling-channel-partners-omdia/822012/
9. Microsoft, "Azure Consumption Commitment benefit," Microsoft Learn: the benefit "only applies to licenses that are exclusively used in Azure"; a license deployed in a hybrid environment or on-premises is not eligible to count toward the commitment. https://learn.microsoft.com/en-us/marketplace/azure-consumption-commitment-benefit
10. Amazon Web Services, "Frequently Asked Questions," AWS Marketplace buyer guide: purchases count toward committed spend (EDP or PPA) "if the products are eligible for AWS committed spend drawdown. Products deployed on AWS typically qualify." https://docs.aws.amazon.com/marketplace/latest/buyerguide/multi-product-solutions-buyer-faq.html

What is not yet sourced

• The Canalys and Omdia figures (reference 1) cover IT spending as a whole, not industrial technology. The 2026 share is cited to Omdia's January 2026 update and to Channel Dive's reporting of its later figure. The marketplace and co-selling figures (references 7 and 8) describe the IT channel, and the Omdia release is read through a syndicated copy.
• Marketplace eligibility for committed cloud spend differs by provider and offer. Microsoft documents that on-premises and hybrid licenses do not count toward an Azure consumption commitment [9], and AWS that products deployed on AWS typically qualify [10]. Other providers' rules were not reviewed.
• Published channel shares vary by vendor, so the paper cites them as examples rather than a market figure: one automation vendor reports about 65% of global sales through independent distributors (Rockwell Automation, FY2025 10-K) [3]; one engineering software vendor about 75% direct (PTC, FY2024 10-K) [4]; another 94% direct and 6% through channel partners (Bentley Systems, 2025 10-K) [5]; one machine vision vendor sells direct to large accounts and through distributors and integrators to smaller ones (Cognex, 2024 10-K) [6].
• The partner type table, the five claims and the seven routes are Thing Company's framing from engagement experience, not an industry standard.
• The observation that partner delivery staff often answer differently from partner leadership rests on Thing Company's experience, as do the statements that integrators often hold the plant relationship and that plant leaders ask them first, that many plants give a mixed answer on who to call, that industrial buyers depend on partners at least as much as IT buyers do, and that the industrial channel is still paid mostly per transaction and per project.

Thing Company is an independent market validation practice for industrial technology. We test whether a buyer exists at a price that works. For technology vendors, that includes testing the partner route with partners and plants before the headcount and incentives are committed.

Planning a partner route into industry? Start with a Sprint Assessment.

Not ready to talk? Run the free partner route test self-check from the toolkit.