Marketplace design optimization that resulted in 30% total revenue per user growth

Summary

We increased the ARPU for the client by 30% by offering users an alternative flow to convert. A user can either post a job and wait for bids from contractors or select specific contractors and offer the job only to them. The first flow is offered by default. The second flow must be selected manually. We gave users an option to select either of the two flows more easily. This move resulted in a 30% growth in ARPU.

Overview

Kabanchik.ua is a two-sided marketplace that bridges users looking to have a specific one-off job completed and contractors who offer their services based on professional proficiency:

1. A user can post a job to the marketplace and wait for contractors to bid for it; 
2. A user can select specific contractors based on reviews and profile description and offer the job to them directly.

Our task was to increase the revenue per user (ARPU) metric by working with the conversion rate (CR) and average order value (AOV).

Research

By looking at the revenue breakdown allocated to each of the two conversion flows, we found that the vast majority of users were choosing the first conversion flow that was offered by default: the flow where a user posts a job on the marketplace and waits for contractors to bid for it. This was the natural flow for users because it let them view bids from multiple active contractors and select the best one.

However, we also conducted user tests and user polls and found that about 20% of new users wanted to know more about the selection of contractors on the marketplace before posting a job. New users wanted to know if contractors are skilled and trustworthy enough to complete their jobs. Ultimately, new users wanted to evaluate if it was worth spending time writing up a job post on this marketplace. 

Solution

Our solution was to combine the two conversion flows into one, and offer users to either post a job and wait for the bids, or to handpick specific contractors and offer them the job based on their reviews and profile. To do that, we introduced the “Popular contractors in this category” block on the form for posting jobs.

Control variation

Alternative variation

We did two things with this solution:

1. We allowed users to explore the selection of active contractors on the marketplace by easily identifying and visiting their profiles. Users used reviews from past jobs to explore contractors’ trustworthiness and professional proficiency. This UX improved users’ trust and motivation to invest time posting a job.

2. We allowed users to easily find contractors that suit their perception of a perfect performer of the job and offer the job to them directly. Based on our logic, this would also increase users’ motivation to invest time posting a job.

Validating the hypothesis

To validate this hypothesis, we conducted an experiment that measured the impact of the new UX on the CR, AOV and ultimately ARPU. The experiment had the following characteristics:

Experiment type: A/B test

Traffic split: 50/50

Key metric: ARPU

Number of users who took part in the experiment: 4,137 users

Number of conversions that were recorded during the experiment: 1,465 conversions