Founding 9Up to 50% off, locked for life·2 claimed · 7 leftClaim your spot
ParrotB
← All three builds
Custom build · Work

The first custom build, in full.

Stefyvet is a mobile veterinary practice in the western Algarve and the first customer to buy all three builds. What was wrong, what was built, and the parts that are honestly not proven yet. One project explained properly is the only evidence worth reading.

The client

Stefyvet, a veterinary practice that goes to the animal instead of the animal coming to it.

Five zones across the western Algarve. Which places are covered, and on which day, is a decision the practice makes and changes. That one constraint shapes everything about how it takes work: an enquiry from the wrong side of the map is not a booking, it is a conversation first.

It is also ParrotB’s first customer, and it joined while the platform was still an alpha.

The project
BuildBrand, Site and Software
SectorVeterinary, mobile practice
LanguagesPortuguese and English, both live
Live atstefyvet.com, and stefyvet.parrotb.com
ModulesBase platform, website, back office. No reception
CoversFive zones: Lagos, Sagres, Aljezur, Monchique, Portimao

No elapsed figure is published for this project. Dates are not a property of the work, they are a property of the calendar: the span from day one to go-live is agreed when a project is scheduled, and it is in the quote before anything starts.

What was wrong

Every enquiry arrived without the one fact that decides whether it can be accepted: where the animal is.

A practice that travels has a shape no ordinary booking form knows about. Some places are covered and some are not, and which day they are covered on is a decision the practice makes rather than a fact a form can assume. An enquiry that does not carry a location has to be answered with a question. An enquiry from outside the area has to be turned down politely, after somebody has already spent time on it.

Before this project there was nothing in front of any of that. No site to send anyone to and no identity to put on one. Work arrived by email and was held together by the platform underneath, which is a workable answer for a first month and not for a business.

What was built

All three builds, which is why this project describes them better than a price list does.

Brand, applied rather than delivered

A mark, a palette, type and the rules for using them. Put onto the live site rather than handed over as a folder, and onto print: a business card, front and reverse. The whole identity exists as design tokens, so it can be applied again later without the designer.

A site whose words she owns

Built on the platform and running on her own domain. It runs in Portuguese and English, both live, edited from one dashboard rather than maintained as two sites. The deliverable is not the pages: it is a locked content schema, the contract that says which fields are hers.

A form builder, and a service area

The platform could not do either before this project. A form that can be sent as a link or sit inside the site, which runs the intake questions a vet would ask, records the animal in a way that stands up to GDPR, and asks where it is. Behind it, service areas drawn on a map and attached to the days they are visited.

The correction that matters, stated plainly

The form builder is live and in use. The service area check is designed and has not shipped. The areas can be drawn and attached to days; the step that reads an address at the point of enquiry and answers yes, Tuesday, or no, not covered, is specified and not yet built. It is described here because it is the clearest example of what a Software build produces, and describing an unbuilt thing as delivered is the one mistake a page like this cannot recover from.

The Stefyvet mark: a veterinary cross holding a cat above a dog, with the practice name set beneath it
The mark, and the lockup it sits in
The business card, front: the mark, the practice name, the word veterinaria and the domain
Card, front
The business card, reverse: name, qualification, phone, email, website and a QR code
Card, reverse. Same card, same size, both sides
The ParrotB dashboard service area screen: five named zones beside a map of the western Algarve with five coloured polygons drawn on it
Five zones, drawn in the dashboard and attached to the days they are visited

What the first customer changed about the product

She did not want the voice assistant, and that refusal is the most valuable thing in this project.

The reception module is the platform’s flagship: an AI that answers the phone. She was happy with everything else and had no use for that one. The obvious response is to sell harder. The useful response is to ask what the module was actually for, which is getting a job into the system without a person typing it.

So intake stopped being a single product and became two. A voice agent for businesses that want one, and a form for businesses that do not, sent as a link or embedded in a site, built by the customer rather than by us.

Both now feed the same place. That is what makes it a capability rather than a favour: the next customer who says the same thing fills in a field instead of commissioning a project.

The same is true of the service area. It began as one practice’s map and became configuration every tenant can use, which is the whole reason a Software build costs what it costs rather than what building it from nothing would cost.

What was measured

The published thresholds, and what this project recorded against them. Measured and reported, not certified.

MeasureThresholdRecorded
Widths390 / 820 / 1440 pxnot recorded
AccessibilityWCAG 2.1 AA, EN 301 549not recorded
Largest contentful paintat or under 2.5 snot recorded
Cumulative layout shiftat or under 0.1not recorded
Interaction to next paintat or under 200 msnot recorded
Why that column is empty, which is the honest version. The go-live report did not exist when this project ran, so the measurements were taken during the work and never written down in a form anyone could keep. That gap is exactly why the report now exists: every build hands over a dated document with these five rows filled in, and the annual check produces the same document again a year later so the two can sit side by side. This project is the reason, not the exception.

No ranking, traffic or revenue claim belongs in that table, and none will be added to it.

The Stefyvet home page at desktop width, the towns covered named above a large headline
Live, at 1440
The same home page on a phone, the headline reflowed and the booking button within thumb reach
The same page, at 390

Live, and honest about the rest

What is not here.

A case study that only lists wins is an advert. These are the things this project has not proven, and they are listed because you are about to spend real money on the strength of it.

  • No outcome numbers. No enquiry counts, no conversion figures, no revenue. The site is live and the measurement period is not long enough to say anything true, so nothing is said.
  • The booking flow is not live. Enquiries arrive through the form. The end-to-end booking, with the area check in front of it, is designed and not shipped.
  • One customer. This is the first custom build, not the tenth. Everything on this page is one project’s evidence and should be read as one project’s evidence.
  • The print and the screen disagree. The site header carries stefaniaamidei.com and the business card carries stefyvet.com. Both resolve. It is a loose end and it is listed rather than tidied out of the screenshot.

What I would do differently

Ask what a customer will not use before designing what they will. The whole project was scoped around the reception module because it is the flagship, and she turned it down in the first conversation. That refusal produced the form builder and reshaped the product for everyone who came after, which was lucky rather than deliberate. It is now the second question in every scoping call: which part of this do you already know you do not want?

Your project would run the same way.

A call, a scope, a fixed quote, a daily email, dated approvals, and a report at the end with those five rows filled in. Fifteen minutes to find out whether it should run at all.

Describe the project

Six questions, then pick a time.

The long version of this project, with the design decisions behind it, is at luislandi.com/stefyvet.