Because the two channels sell to different buyers with different decision processes: a catalog buyer wants a stocked, standardized, single-unit item delivered tomorrow, while an institutional buyer wants a configured system with installation and training, and each channel rewrites what the factory behind it has to be good at.
The catalog channel turns a manufacturer into a master-data and logistics business. A distributor fills a catalog page with breadth rather than explanation, so the factory must supply a broad set of SKUs that is ready to list instead of a small set of projects. That means make-to-stock production with predictable lead times, packaging built to survive a parcel network when a single unit ships alone, barcoded cartons, and a complete master-data record per item: description, unit of measure, weight and dimensions, images, safety documentation, country of origin and the customs code used on the export documents. It also means documentation written for a non-specialist buyer, because the person reading the manual may be a biology teacher with a limited budget and no spectroscopy training, plus a spare-parts list short enough that a dealer can quote it from a desk. The industrial consequence is tooling fixed for years, long production runs, disciplined engineering-change control so that one revision does not invalidate thousands of printed manuals, and profit that comes from repeat volume on a stable catalog rather than from any single order.
Direct institutional selling replaces catalog breadth with configuration and support. A university or a ministry of education buys through tenders and framework agreements with published technical specifications, and it expects a compliant response on every line, custom configuration, delivery and installation, teacher training, and warranty with spare parts over a defined support period. Sales cycles are long, order counts are low, order sizes are large, and revenue arrives in budget-shaped lumps rather than steady weekly shipments. A factory serving that buyer needs engineer-to-order capability, staff who can answer a two-hundred-line specification, project management for installations that must fit around term dates, field service either employed or through certified partners, and a spare-part commitment that will outlive the person who signed the purchase order. A maker optimized for the catalog loses tenders because it cannot answer the specification or schedule the installation; a maker optimized for tenders fails in the catalog because it cannot hold stock, and its price carries services the distributor never asked for.
Most global suppliers therefore run two businesses inside one company, and the buyer should read the channel rather than the brand. The useful questions are concrete: was the unit shipped through a dealer built to the same drawing and the same end-of-line test as the unit sold direct, who owns the warranty obligation in year seven, does the dealer stock the item or drop-ship it from the plant, which legal entity issues the invoice and signs the EU declaration of conformity, and who is the named contact when a class set of instruments arrives with two units out of calibration. Channel choice does not determine quality, but it does determine lead time, price structure and who answers the telephone a decade after the equipment entered the school.
Because education is a fragmented, budget-cycle-bound slice of a portfolio whose turnover is driven by industrial, clinical and research customers, which gives a school real manufacturing strength but also exposes it to portfolio decisions taken for somebody else.
Why the education line is structurally small. Teaching demand is spread across tens of thousands of schools and colleges, each buying a few units, and the person choosing is usually a teacher or a department head spending a limited budget rather than a capital committee approving a fleet. Unit prices are low against research or industrial instruments, replacement cycles run for many years, and purchases cluster into fiscal windows, so revenue is lumpy and hard to forecast. Group turnover, by contrast, is built on analytical, clinical and industrial customers who buy instruments with service contracts attached. That is why a group well known in schools still reports education as a small share of what it sells, and why an optics manufacturer can build teaching microscopes on the same optical and mechanical base as its industrial lines without education driving the business at all. The education catalog sits beside the instrument business rather than under it.
What the buyer gains from buying into that structure. The teaching product is usually produced in a genuine plant with real process capability: optical fabrication and coating, surface-mount assembly, injection moulding of housings, and in-house calibration and end-of-line test fixtures rather than a bench borrowed for the purpose. Group purchasing power on sensors, optics, power supplies and displays means the teaching line inherits parts that were qualified for other markets, and the group's incoming inspection, tolerancing and change control are typically stricter than those of a small assembler. Certification usually covers a named site and legal entity, which makes verification straightforward, and a group that already supports industrial instruments for decades is organizationally capable of keeping a spare-part line open for a long time. For a university buying twenty identical instruments for a teaching laboratory, that factory discipline is worth more than any marketing claim.
What it exposes the buyer to, and how to reduce it. The education SKU competes for engineering attention and factory capacity against programs that are orders of magnitude larger, and it rarely wins. A component discontinued for the group's benefit elsewhere can force a redesign of a teaching product whose schools expected to run it for a decade, a portfolio review can retire a line quietly, and a service organization measured on contract value has limited incentive to expedite a single classroom replacement part. The mitigations are contractual rather than technical, and a buyer can ask for them: a written spare-parts availability commitment over the expected service life, notice terms if the group decides on a last-time buy, confirmation that the teaching product uses widely available consumables rather than proprietary parts, a stated roadmap commitment, and the name of the legal entity and site responsible for the product. Buying from a diversified group is not a mistake; buying without checking where the education line stands inside that group is.
A bench-scale process trainer has to reproduce the causal relationships a student is meant to learn, namely what a sensor sees, how an actuator moves the process and how a fault spreads downstream, while deliberately abandoning the scale, the fluids, the pressures and the stored energy that make the real plant dangerous.
What must be preserved faithfully. A rig is a model of behaviour, not of appearance. Loop behaviour has to survive: gain, lag, dead time, and whether the controlled variable is self-regulating or integrating, because a student who tunes a level loop on a trainer and then meets an integrating process on site has learned the wrong reflex. Sensor placement must keep its relationship to the actuator and the disturbance, so a student can see why a temperature element close to a heater reacts before the one in the vessel body. Actuator response must be honest: valve travel time, trim characteristic, deadband and resolution decide whether a loop is limited by the process or by the final control element. Fault propagation matters just as much, because diagnosis is the skill being taught: what happens downstream when a control valve sticks, a sensor drifts, a line plugs or a transmitter fails high. Dynamic similarity, not visual similarity, is the design target, and a manufacturer can demonstrate it with measured step responses rather than with a photograph.
Where fidelity is deliberately given up, and why that is harder than it sounds. Real process fluids become water, air or another benign surrogate; pressures and temperatures fall to bench-safe levels; hazardous energy is removed, so plant three-phase supplies, steam headers, production-scale rotating machinery and stored hydraulic energy give way to low-voltage, guarded, interlocked subassemblies; vessel volumes and pipe runs shrink so residence times and thermal inertia no longer match the plant. The trap is that scaling is never neutral. Lowering temperature changes the driving force for heat transfer and moves sensor time constants, so a temperature loop that is sluggish in the plant can look fast on the bench. Shrinking a vessel changes mixing and dead volume, so a reaction that is mixing-limited in production may appear kinetically limited on the rig. Substituting gas for steam changes compressibility. A training rig is therefore not a small plant; it is a different plant tuned to teach the same control lessons, and vendors who understand their own product say so.
Features that exist for teaching and have no counterpart on real plant. Fault-insertion switches let an instructor create a failure on demand; transparent sections and cutaways let a student see internals that would be behind steel plate on site; instrumented tie-points expose signals an operator would never be given; guards and interlocks cover hazards that real plant manages by permit; and a start-up or trip sequence that takes hours in production runs in minutes on the bench. Two consequences follow. A fault switch produces a fault that is pre-defined and repeatable, while a real plant fails in unplanned combinations, so the rig teaches diagnostic technique rather than plant history. The rig's own safety case is real too: teaching equipment sold in Europe still needs CE marking under the applicable EU directives, including the machinery regulation that replaces the Machinery Directive, plus the electromagnetic compatibility and low-voltage rules for its electronics, all supported by a technical file. Ask for the risk assessment and the declaration of conformity, not only the certificate.
They compete on the parts of the contract a global group treats as overhead, namely currency, freight, duty, local-content eligibility, service reach and spare-part logistics, and they lose whenever a tender rewards brand, software ecosystem or global service coverage instead.
Tender rules and local content decide more outcomes than product quality does. Public procurement commonly runs through framework agreements and tenders with published technical specifications and weighted evaluation, and many jurisdictions add a domestic-manufacture or local-content element. India's public procurement preference framework classifies bidders into tiers by measured local content, which affects eligibility and price preference, so a supplier that imports finished equipment and re-badges it can fail to qualify regardless of the address on its letterhead. China's government procurement rules likewise favour domestic products in many categories. European Union member states must apply single-market rules, so domestic weighting there is more often expressed through the tender's own scoring, delivery terms, documentation language and support requirements than through an explicit nationality test. Localisation is substantiated with cost accounting, certificates of origin and manufacturing records, and the rules vary by jurisdiction and by the specific tender, so the honest reading of a local-content claim is that it describes who performed the value-add in that jurisdiction, not where a company is registered.
Price structure is a currency and logistics question before it is a labour-cost question. A regional manufacturer in Spain, India or China carries engineering, machining, assembly and test costs in its own currency, ships short distances, and avoids import duty and customs clearance delays on finished goods, so it can quote a complete tender without the margin stack of a distributor plus an integrator. That does not make it cheaper in every line item. A global group buys sensors, optics, displays and power supplies at a volume no regional maker can match, so the gap narrows on electronics-dense items and widens on sheet-metal enclosures, machined parts, frames and benches, where local fabrication is decisive. Exchange-rate movement can invert the comparison over the life of a framework agreement, so a quoted price is partly a bet on the currency. Volume cuts the other way too: a factory building two hundred benches a year cannot amortize injection moulds the way a group can.
Service reach and spare-part logistics are where regional manufacturers are genuinely stronger, and where the risk moves to the buyer. For a school, the binding constraint is rarely the purchase price; it is how long a laboratory sits unusable. A regional manufacturer can place a technician and a stock of spares within a day's road travel of its home market and hold slow-moving parts for an installed base it knows by name, while a global group routes the same request through an appointed partner whose engineer may cover several product families and several countries. The trade-off is balance-sheet depth: a local supplier's ten-year spare-part promise is worth what that company can support over ten years. The answer is to demand the same evidence from both: a named legal entity, a parts list with prices and lead times, a stated response time, a calibration and repair route, and a support commitment written into the contract over a defined period, rather than assuming that the global brand is automatically safer or the local maker automatically faster.
The claim is verifiable, because manufacturing leaves documents behind: certificates with a defined scope and a named site, tooling and test-fixture ownership, process records, and a plant that a customer is allowed to walk through.
Read the certificate, not the logo. Ask for the certificate number, the issuing certification body and the accreditation behind it, then verify it on the issuing body's register or in an international database of accredited management-system certificates such as IAF CertSearch. The detail that matters is the scope text: a certificate limited to design, distribution or trading is not evidence of production, and a scope that names one site says nothing about a second plant shipping your order. Check that the certified legal entity and address match the party answering the tender and the party that will issue the invoice. The relevant schemes are ISO 9001 for quality management, ISO 14001 for environmental management and ISO 45001 for occupational health and safety; IATF 16949 is the automotive quality standard and is not required for teaching equipment, but a supplier that also holds it for a sister plant usually has stronger process discipline. Where calibration is involved, ISO/IEC 17025 says a laboratory is competent to issue a calibration certificate you may one day rely on.
Trace the value-add rather than the brand. A genuine manufacturer answers with equipment, capacity and names: who populates the printed circuit boards, and whether the surface-mount line, reflow oven and automated optical inspection equipment are in the same building or at a contract assembler; who cuts, forms and coats the sheet metal; who machines the parts and on what machines; who owns the injection moulds and where they are stored; who owns the calibration and end-of-line test fixtures, and whether those fixtures trace to a national metrology institute. Ask for a routing on one representative product and for a capacity figure in units per month. Ask where rejected and returned units are reworked, and whether a serial-number prefix identifies a plant, because a unit that goes back to a third country for repair was probably not made in the one you visited. A relabeller answers these questions with a catalogue page, a brand story and a photograph of a building; a factory answers them with a list of machines and the name of the production manager.
Test the claim with access, paperwork and the small signals. Ask for a factory visit and an audit, and note whether the reply is a proposed date or a reason why visits are inconvenient. Ask for the EU declaration of conformity and the technical file for a CE-marked product: the declaration names a legal entity and an address, and a sales office with no production, or a declared manufacturer different from the entity bidding, is a discrepancy worth resolving in writing before award. Ask for a bill of materials on a single assembly and which critical parts have an approved second source. Ask for the country-of-origin declaration and the customs code used on the shipping documents, since origin is a statement about where substantial transformation occurred rather than where the carton was closed. Then weigh the physical signals: an operating site has goods-in inspection, work in progress, a test bay, a fixture store and a spare-part store, and a company that manufactures tends to market capability, while one that assembles and relabels tends to market brand, awards and the factory of a partner.
