Mybuilder.com

Web / Mobile

Project Overview

Mybuilder.com are a popular marketplace service that connect homeowners with tradespeople across Europe.

This project involved the design of chat assistant feature to enable improved marketplace communication between homewowner and tradesperson

My responsibilities

Senior Designer

My responsibilities as lead designer involved setting strategic design direction, full stack UI/UX, setting a/b tests, and user research.
‍
Sept 2023 — Jan 2024

Design hypothesis

Research suggested that tradespeople do not know when best to suggest home visits to quote on listed jobs. Chat between the two parties then sits stagnant and the job is often lost.

By introducing 3rd party chat facilitation messages (as mybuilder.com) between homeowners and tradespersons, conversations will happen at the relevant time in the process home visits and job conversions will increase.

We introduced controlled a/b testing which sent facilitated messages at different points to test which time periods converted a final job acceptance better.

** A homeowner will 'hire' a tradesperson. Confirmation of this is classed as a conversion**

** A Tradesperson will pay to send an approach to the homeowner so it is essential that conversations often convert to hires**

Research questions

It was vital to ask specific questions as regards the behaviour of the tradesperson and homeowner through the job process. Interviews and research were able to answer the following design questions:

What is the average time period between initial contact from the hownowner to a hire?

What is the average time of completing a specific job from initial contact?

How long are specific jobs classed as 'in progress' (before a completion comfirmation)?

What percentage of initial contact conversations do not result in a hire?

How many messages are sent on average for successful job hire?

Design problem - No idea of when to faciliatate

We had no idea of when to send facilitation messages and at what stages of the process tradespeople were looking to quote and invoice homeowners. Research was to provide the logic for this.

Design problem - Conversation sits dormant

Chat is currently static and does not prompt specific actions at logical points in order to facilitate and encourage conversion. Data showed that jobs that sat dormant for more than a few days rarely converted.

Design overview

When a homeowner shortlists a tradesperson, a chat opportunity is created. Reseach showed chat's often sat dormant with neither party seeing it as their responsibility to send the first message, or, chase initial messages.

By introducing each party and displaying the others intentions, it encouraged more opening communication.

Controlled tests showed a higher rate of opening messages are sent after a facilited introduction.

Introduction messages

Design overview

Research showed that job enquiries that resulted in a home visit to quote, often had higher conversion rates - Home visis were a core metric to uplift.

Home visits often didn't happen as tradespeople didn't suggest or enquire about availability at the correct point in the conversation (often missing the opportunity by +2/3 days)

Chat facilitation prompted the idea of a site visit when the chat had reached 48 hours without a home visit being manually requested/scheduled by the tradesperson.

Controlled a/b tests showed more home visits occured with an automated home visit facilitation messsage, than without.

Prompting a home visit

Design approach

A core problem with the model of the Mybuilder marketplace was that conversation was often taken offline after the initial introduction and data was hard to gather in terms of job completion rates.

Asking the tradesperson what the status of the job was after 7 days of chat inactivity allowed for much better progress tracking and the ability to push for the request of reviews - a vital part of a tradespersons portfolio on the marketplace.

Tradespeople were prompted again 24 hours after the initial 7 day prompt period.

Checking on job status

Design approach

Research showed that a detialed reviews section on a tradespersons portfolio much improved the chances of winning and converting a job.

After confirmation was received on the job completion, it was prompted that the tradesperson requested a review of their services from the homeowner.

Without this prompt, testing showed that reviews of completed jobs were missed. Stronger reviews equalled more hires and increased reveue.

Controlled a/b tests showed more reviews were requested when sending an automated review prompt.

Prompting for a review

✅ Prompting job compeltion = More avialble data

Prompting for a job completion status helped with metric tracking and job understanding

✅ Introduction messages = more converstiaon

A higher rate of messages are sent after a facilited introduction which leads to a greater chance of a quote being requested

✅ Prompting reviews = Stonger conversions chances

Prompting the Tradesperson to request a review resulted in stonger portfolios and higher conversion rates long term

✅ Prompting home visits = higher quote success

More home visits occured with an automated home visit prompt messsage, than without.

So, why not take a look at my next case study?