After you've made a simple GET request (see Quick Start), and tried a few List request, you can send a Crowdsourcing Report. This is a way for end users to send feedback on MRS results.

Examples that your users may wish to report:

  • Incorrect merchant ("This is Jacob's Bakery not Jacob's Mechanics")

  • Incorrect location ("This is not in Bristol, it is in London")

  • Incorrect logos ("This is the logo for Mandy's Flowers not Mandy's Pizza")

As a goal, it continually improves the clarity and accuracy of the information and further reduces the calls to the call centre + match rates + user trust.

Crowdsourcing is listed as a separate step because it requires extra user interface development (e.g. prompt to fix/report error at bottom on transaction details).

Here's an example of how you might implement a crowdsourcing sequence:

  1. Your app enriches a transaction, using the /merchant/list or /merchant endpoints (see Quick Start). It records the X-Request-Id and X-Timestamp headers that are returned for each enriched transaction. It also records the transactionId.

  2. This enriched transaction is shown to the end user in your banking app.

  3. The user notices that the merchant name is wrong - it's Peter's Pizza, not Peter's Pezza!

  4. They click on the transaction in your app, taking them to a 'Detail' screen (see Recommended UX).

  5. The Detail screen has a button - 'Improve name, location or logo'. The user clicks this button.

  6. They specify the correct name in a dialog - "Peter's Pizza", and click 'Send'.

  7. Your app takes the X-Request-Id and X-Timestamp headers that were originally recorded for that transaction and sends them to the /crowdsourcing endpoint, along with the User's Feedback.

Next, let's see how to implement this in Crowdsourcing Example.

Last updated