How nano Scales Without a Support Team: A Look Behind the Scenes

nano sells toothbrushes every day, answers customer questions and handles returns. And nobody is employed to do it. Scaling e-commerce without employees sounds like a LinkedIn fairy tale, so here is the look behind the scenes.

nano is our own brand. No client project, no glossy case study. I will show you which processes really run automatically, what we humans still do ourselves, and what went wrong along the way.

The starting point

nano is one of three brands Damian and I run, alongside mate and MUSTAX. Damian described how that works with two people in general in how we manage 3 brands with 2 people.

nano takes in 50 to 150 orders a day. On top of that come around 150 to 200 customer requests a week across all brands. Calculated the classic way, that would be at least one full-time role for support and operations, more likely two.

The alternative to hiring was never "work more ourselves" for us. The first two years were exactly that, and they almost burned us out. The alternative was to build processes that run without people.

What runs automatically at nano

Order handling and fulfilment

The biggest chunk. Order in Shopify, handover to the fulfilment partner, tracking back, email to the customer: that runs completely without us, 24/7. It used to be 4 hours a day, today it is 15 minutes, and only for edge cases. The complete technical setup is in order processing without staff.

Customer service, the biggest part of it

"Where is my parcel?" makes up around 40 percent of all our requests. The AI pulls the tracking info and answers by itself, in minutes instead of hours. In total about 65 percent of requests get answered entirely without us.

The rest gets categorised and pre-drafted. We review and send. The rules behind that, and what we deliberately keep away from the AI, are covered in detail in our guide to customer service automation.

Returns

The customer fills in a form, the label arrives automatically, once the goods are received the refund runs by itself. Only "damaged" and "other reason" land with us. At mate this system took us from 45 minutes to 5 minutes a day, and nano runs the same setup.

Reporting and bookkeeping

The numbers come to us automatically, we do not go looking for them. Revenue, return rate, anomalies: it all lands in Slack. Bookkeeping also runs largely automatically, from the bank transaction to the DATEV export for our tax advisor, see our bookkeeping setup.

What humans still do

Now the part the "passive income" bros leave out. nano does not run without people. It runs without employed people.

• Product and strategy. Which products, which prices, which direction: we decide that, not a workflow.
• Marketing decisions. The execution is often automated, the concepts are not.
• Emotional customer cases. An angry customer whose gift never arrived gets a human. Always. The AI escalates cases like that to us automatically, and that is one of our most important rules.
• Product advice with depth. Questions like which brush suits sensitive gums we answer ourselves. It builds trust, and that is exactly where a brand comes from.
• The automations themselves. Somebody has to build, monitor and improve the systems. Automation means supervising instead of executing, do not forget that.

Put differently: we did not automate a support job away. We prevented it from coming into existence.

The learnings, including the uncomfortable ones

1. Automation is a filter, and it replaces nobody

The system keeps repetitive requests away from us so we have time for real problems. Use automation as an excuse for bad service and you lose customers. Our satisfaction scores got better after automation rather than worse, precisely because response times dropped.

2. The build took years, not weeks

We did not build an autopilot over a weekend. First came the classic hustle, then process by process the automation. Anyone promising you that your shop will run without you in four weeks is selling you something.

3. Error handling is the real work

The happy path is built quickly. Most of the development time went into one question: what happens when something goes wrong? Every error creates a Slack alert here, and after several failed attempts the case moves to a manual list. Without that net the system would be negligent.

4. The model has a growth limit

Without a team you scale processes, and not necessarily everything else. New markets, retail listings, big product launches eventually need hands and heads. Automation pushes the point where you have to hire far into the future. It does not remove it.

5. Peaks are the proof

The moment it all paid off: Black Friday. nano ran a multiple of normal volume with the same effort on our side. Manual processes scale linearly, automated ones practically stop scaling at all. That is the whole point.

Who this model works for

Scaling e-commerce without employees does not work for everyone. The frame for it:

Good conditions:

• Product brand with standardised, repeatable processes
• Enough volume for automation to pay off
• Willingness to put time into the build up front

Bad conditions:

• Heavily customised products or a consulting business
• No technical understanding in the team and no partner for it
• The expectation that it runs without supervision

Which processes to tackle first is in our overview of e-commerce process automation.

Conclusion

nano scales without a support team because machines handle the standard cases and we take on the exceptions. That is no autopilot and no magic. It is the result of years of systematic work, with monitoring, escalation rules and clear limits for the AI.

The gain is more than time. The gain is that growth stops being scary, because more orders no longer mean more work.

If you want to build your business in that direction: that is exactly what we help brands with at Flowhouse. We are happy to show you what the setup behind nano looks like.