We use analytics and a chat widget to understand and support visitors to this site. They only load if you say yes. What we use.

How teams actually run this

Not case studies written backwards from a number. What these businesses send, what comes back, and what happens the moment somebody is unhappy.

The one that matters

Reaching the person who gave you a 4

A score on a dashboard is a number. A score attached to an email address, a comment and a timestamp, delivered to the place your team already works, is a conversation you can still have. The businesses getting the most out of Downright all built the same thing, and none of them wrote much code to do it.

  1. The survey goes out from their own system

    Not from a marketing calendar. An order is delivered, a wash is collected, a ticket is closed — and that event calls POST /api/v1/surveys/:id/invitations with the customer's email. The survey arrives while the experience is still fresh, which is most of why the response rates hold up.

  2. The answer comes back as a webhook, not a report

    The moment somebody answers, Downright posts a response.created event to whatever URL you gave it. It carries the score, the comment in the customer's own words, the survey it belongs to, and — unless the answer was anonymous — who they are: 

    {
  "event": "response.created",
  "survey":   { "id": "…", "name": "Post-delivery", "question": "NPS", "channel": "api" },
  "response": { "id": "…", "score": 4, "comment": "Arrived late and no one told me",
                "answered_at": "2026-08-28T09:14:22Z" },
  "contact":  { "email": "someone@example.com", "name": "Someone" },
  "account":  { "name": "Your Brand" },
  "sent_at":  "2026-08-28T09:14:23Z"
}

  3. A scenario decides what to do with it

    This is where Make.com, Zapier or n8n earn their keep, and where nobody has to write a service. The webhook lands in a scenario, a filter asks one question — score <= 6 — and only the answers that need a human get past it. Everything else is counted and left alone.

    What teams do on the other side of that filter is always some version of the same thing: post it to the Slack channel the shift is watching, open a ticket in Zendesk or HubSpot with the comment already in the body, add a row to the sheet the manager reads on Monday, or send the branch a message naming the customer to call back today.

  4. Somebody gets in touch the same day

    That is the whole trick. Nothing here is clever — the value is that the gap between "a customer was unhappy" and "somebody knows" is minutes rather than the end of the month, and that it happens without anyone remembering to check a dashboard.

Every webhook is signed, so your scenario can verify Downright sent it. Setup, payloads and error codes are in the API reference.

The businesses doing it

Two shapes of problem that look nothing alike, asking the same two questions.

Foodcircle

B2B marketplace for food ingredients

A marketplace has two sides and only one of them is usually asked. Foodcircle sells food ingredients between businesses, where a single order is large, the relationship is long, and a bad experience is quiet: buyers rarely complain, they just buy somewhere else next quarter.

That is the case NPS was built for, and the reason it is worth automating. The survey goes out from the systems that already know an order happened, rather than from a marketing calendar that does not. The score gets tracked over time and the open question underneath it is where the reason lives — in the customer's own words, which is the part a team can act on.

Downright has significantly streamlined our customer feedback process. The automated NPS surveys are easy to set up and provide valuable insights. It's a must-have for any growing tech startup! Maria M., Foodcircle

Washr

Laundromat chain, South Africa

A laundromat has no email list. The customer arrives, uses a machine, and leaves, and in most chains that is the entire record of them. Asking anything at all means asking in the room.

That is what the link channel is for: one public URL, behind a QR code on the machine or the door, that anybody can answer without being in a database first. The answer is anonymous unless somebody volunteers who they are, and it still carries the one thing worth having — the reason, written by the person who has just used the place.

For our laundromats, the introduction of automated NPS surveys through this tool has been revolutionary. It's boosted our response rates and provided us with immediate insights that are crucial for enhancing customer experience. David D., Washr

Four ways the question reaches somebody

Sending is unlimited on every plan, so the channel is a question of what fits the business, never of what it costs.

The API

Your system decides who gets asked and when. One POST per person, triggered by whatever event means "this customer has just experienced us".

Email

Paste a list or upload a CSV and Downright sends and throttles it, so nobody is surveyed twice in a quarter.

A link or a QR code

One public URL anyone can answer. On a receipt, on a door, on the machine. Nobody has to be in your database first.

Embedded

The scale sits inside your own page or app, and the answer is recorded without the customer going anywhere.

Free plan, no credit card. The API and webhooks are on it too.

Get started Read the API docs