The promotional products industry is dividing into two order types: high-volume, low-value work that must be instant and automated, and high-value, complex projects that need human expertise. Serving both with one process erodes margin. The fix is a two-engine operating model, one tuned for speed and one for service, with a clear handoff between them.
Two orders land in your inbox on the same morning. The first is 250 branded pens for a returning customer who reorders every quarter, same logo, same colours, same shipping address. The second is a full merchandise programme for a new client: custom packaging, four decoration methods, a colour-matched brand palette, and a launch date that cannot move. Your team runs both through the same process. The same people, the same proofing steps, the same email threads. One of those orders makes you money. The other quietly loses it. And the uncomfortable part is that you probably cannot tell which is which until the month closes.
This is the story of almost every promotional products business right now. The market is separating into two very different kinds of work, and most companies are still trying to serve both with one set of processes. That middle ground is exactly where margin goes to die.
The promo industry is quietly splitting in two, and most companies are stuck in the middle. On one side you have high-volume, low-value orders that need to be instant, automated, and scalable. This is the world of reorders, small quantities, standard products, and simple decoration. Buyers here expect an ecommerce experience: upload a logo, see a proof in seconds, place the order, done. In this lane, automated artwork flows, instant proofing, and production efficiency decide your margins. Speed is not a nice-to-have. It is the product.
On the other side you have high-value, complex projects that require real human expertise and service. Brand programmes, new product development, unusual materials, multi-method decoration, tight timelines, and demanding stakeholders. These orders are won and kept by people, not by portals. The buyer is paying for judgement, reassurance, and a partner who can solve problems that do not fit a template.
Both types of work are legitimate. Both can be profitable. The problem is that they reward completely opposite operating models, and the average company runs a single process that is tuned for neither.
When you serve both lanes with the same process, you get the worst of both worlds. Manual work kills your low-value orders. The moment a human has to touch a 40-euro reorder, redraw a logo, chase an approval, or rekey a spec, the margin is gone. The order was never big enough to absorb an hour of skilled labour, but that is exactly what your one-size process spends on it.
At the same time, over-automation kills your high-value relationships. When a client running a six-figure programme gets pushed through the same self-service funnel as a pen reorder, they feel it. The thing they are paying for, expertise and attention, is precisely what the automated path strips out. You win the efficiency and lose the account.
A single process cannot be optimised for both instant throughput and deep service at the same time. Every compromise you make to speed up the complex work makes the simple work slower, and every safeguard you add for the complex work adds friction to the simple work. The result is an 'average' process that is too slow and expensive for your cheap orders and too rigid and impersonal for your expensive ones. That is not a strategy. That is a margin leak with a workflow diagram.
It is tempting to read the split and conclude that you have to choose: become a lean automated fulfiller, or become a high-touch creative agency. That is the wrong lesson. It is not about choosing one lane. It is about designing two completely different engines inside the same business. One engine optimised for speed and automation. One engine optimised for creativity and service. Same company, same brand, two deliberately different ways of moving work through the door.
The efficiency engine exists to move high-volume, low-value orders through with as close to zero human touch as possible. Its job is throughput and consistency. The design principle is simple: a human should only ever touch an order in this engine by exception, never by default.
Concretely, that means the artwork problem has to be solved by software, because artwork is where these orders silently stall. In our own FastEditor Artwork Automation Benchmark across 13,773 real logo uploads, about 85% of the files customers uploaded were not print-ready. Wrong format, low resolution, raster where you need vector, unmatched colours. If a person has to fix each of those by hand, your low-value lane is dead on arrival. The efficiency engine leans on automated vectorization to turn a customer logo into a clean production file in seconds, instant proofing so the buyer approves without a back-and-forth, and production-ready file generation so what the customer approves is exactly what the machine prints. Wrap that in a product configurator and the order places itself.
The metrics that matter for this engine are cycle time, touch rate (what share of orders needed a human), automated proof approval rate, and rework. If those numbers are healthy, the engine is doing its job, and the small orders you used to lose money on start contributing again.
The expertise engine exists to win and deliver high-value, complex projects. Here the scarce resource is not machine time, it is your best people. So the design principle inverts: automate everything around the expert so that their hours go to judgement, creativity, and relationship, not to admin.
The expertise engine still uses automation, but as a tool in the expert's hands rather than a replacement for them. A specialist should be spending time on decoration strategy, material choices, and reassuring a nervous client before a launch, not manually preparing files or building proofs from scratch. Give them fast tooling in the background and their capacity for the work that actually justifies the margin goes up.
The metrics here are completely different: win rate on complex bids, project margin, client retention and reorder value, and net revenue per specialist. Cycle time barely matters. A programme that takes three weeks and delights the client is a win. The same three weeks spent on a pen reorder is a disaster. Same clock, opposite meaning, which is exactly why the two engines cannot share one process.
Once you accept the two-engine model, the real challenge is not building either engine. It is deciding, order by order, which engine a piece of work belongs in, and moving it cleanly between them. As one reader put it after the original post, the challenge is not simply deciding what to automate, it is recognising when a customer or project needs to move from the efficiency engine to the expertise engine. Companies that define that handoff well protect both their margins and their relationships.
Get the handoff wrong in one direction and you route a complex, high-stakes programme through the self-service funnel, where it breaks and takes the client's trust with it. Get it wrong in the other direction and you burn a specialist's afternoon on an order the software should have closed in ninety seconds. The handoff is not a nice detail. It is the control system for the whole business.
The good news is that most orders sort themselves cleanly if you write down the rules. A useful starting filter:
Turn those five questions into an explicit routing rule at intake, and most of your orders will land in the right engine automatically. The remainder, the genuine edge cases, are exactly the ones worth a human decision.
You do not have to rebuild the company overnight. The move to a two-engine model is a sequence, and it can start this quarter.
The two-engine logic shows up differently depending on where you sit, but no one escapes it. Suppliers and decorators feel it hardest in production: the pressure to accept small, complex, personalised orders at speed while protecting the economics of the line. This is where connecting products once and letting them flow automatically to every channel matters, so that a configuration built for one reseller becomes available across the network without manual rework. Resellers and distributors feel it in the buying experience: transactional customers want an instant, self-service online design and ordering path, while key accounts want a named human. Trying to give everyone the same experience means overspending on the small accounts and underserving the big ones.
Across all of them, the companies that figure this out early are already creating a serious gap. They are winning the low-value work on cost and the high-value work on service, at the same time, because they stopped forcing both through one process. The rest will keep feeling constant pressure on both margins and customer experience, and will not fully understand why, because from the inside a single 'efficient enough' process always looks reasonable.
Not necessarily two teams, and rarely two disconnected systems. It means two clearly different workflows with different rules, different success metrics, and a defined handoff between them. Many companies run both engines on shared infrastructure, with automation doing the heavy lifting in the efficiency engine and supporting the experts in the expertise engine. The separation that matters is in process and measurement, not necessarily in headcount.
Usually the efficiency engine, because that is where the hidden losses are largest and the fix is most mechanical. Automating artwork preparation and proofing on your high-volume, low-value orders tends to recover margin fast and frees up the very people you need for the expertise engine. Once the simple work runs itself, you have the capacity to invest in the complex work.
Only if you apply it to the wrong orders. Automation damages relationships when you push a high-value, high-touch client through a self-service funnel. Applied to the low-value lane, it does the opposite: fast, error-free proofing and ordering is a better experience for a transactional buyer than a slow manual back-and-forth. The relationship risk is a routing problem, not an automation problem.
Two quick signals. First, your skilled people regularly spend time on small orders that cannot possibly justify their hourly cost. Second, your biggest, most strategic clients occasionally get handled by the same generic process as a one-off reorder. If either is happening, you are running one lane where you need two engines.
