Using goals to enourage account linking and increase earnings

Interventions
In-app goal feature
Experiment Type
Field Experiment
Goals
Access a service
Outcomes
Increase earnings
Focus Areas
Product
Behavioral Concepts
Salience Goal-gradient theory
Partner
Steady
Partner Type
Fintech/tech
Collaborator
Irrational Labs

What Happened

It worked. Although the goal feature did not increase the number of app users who linked their bank accounts, it did lead to a significant increase in how much these app users earned through additional jobs in the gig economy.

Lessons Learned

The addition of an in-app goal feature can lead to significant changes in the income generated by users. The ease of scalability of such a feature leads to the potential of large-scale impact.

Background

A recent salary report by Indeed found that 81% of workers in the US don’t feel like they’re earning enough. When asked how much more it would take for them to feel good about their earnings, the majority of participants said they’d need to make at least $6,000 more per year, or just $500 more per month.

One of Steady’s core intentions is to help gig workers attain their financial goals, which frequently includes increasing income; on average, users have earned more than $5,500 with Steady. To help in the most effective manner possible, Steady needs users to link their bank account to the app.

Together with Steady, we designed an experiment with the objectives of 1) boosting the number of users who linked their bank account to the app; and 2) increasing Steady user income.

Key Insights

Our two primary questions for this experiment:

  • Is it more effective to ask members to link their account before or after they set a Goal?

  • Do Goals motivate people to work more and make more money?

In the experiment, we tested the impact of a Goal feature vs. a historical monthly income tracker. We also tested whether prompting people to link their bank accounts first or set their income goals first maximized bank account linking. There were three conditions:

  1. Control: Income Tracker (existing feature) + Link Bank Account First

  2. Goals Feature: Link Bank Account First

  3. Goals Feature: Set Income Goal First

Key insights for the link-bank-account experiment

Link account first conditions:

  • Making the link account salient could maximize bank account links.

  • Bank account linking feels incomplete.

  • The obfuscated feature behind the CTA creates curiosity about the benefit (feature).

Set-goals-first condition:

  • This utilizes the “foot in the door” technique, which is to make a small request (set an income goal) followed by a larger request (link your bank account).

  • Provides a logical reason + product benefit for linking a bank account.

Key insights for the goals feature experiment

  • The feature is highly salient.

  • Goal setting and goal gradient theory could lead to additional effort and income.

  • A weekly goal is the time frame we believed most immediate and easiest to track and remember for people looking for gig-type work.

  • We use implementation intentions to strengthen the impact of the goal feature, prompting a user to establish how they will reach their weekly goal.

Experiment

There were three conditions in the experiment:

  1. Link Account > Income Tracker (Control)

  2. Link Account > Set Goal > Goals Feature

  3. Set Goal > Link Account > Goals Feature

For users in all conditions, push messages designed to encourage goal-setting and plan-making (while remaining congruent with the Income Tracker) were sent every Monday and Wednesday at 11:35 local time. All conditions received the same messages. Examples:

  • Monday: New week, new chance to earn! Check your 💰 from last week & plan how to meet your income goals this week ☑

  • Wednesday! Half-way through the week ☑️You got this! Check your 💰 vs. your goals & handle your business! ✊

Results

Account-linking experiment

Surprisingly, the control resulted in the highest bank account link rate (p<0.014). There was no difference between the Goals variants in terms of bank account linking. Having users input their Goal first did not boost the number of users who linked their bank account.

One possible reason for this unexpected effect is the visual of the rising line on the graph on the “view income” screen, which may make it attractive. It’s also possible that returning users recognized the graph, and its congruence with their experience made it compelling.

Goals Feature Experiment

The Goals feature clearly did increase user income. Users who linked their bank accounts in the Goals feature conditions made more money. In fact, having a Goals feature boosted user income by an impressive additional $7-20 per week (2-6%). The 6% effect represents the impact for users of all incomes over the entire experimental period, the 2.3% effect removes potentially outlying incomes (1.1% of data) across the entire experimental period, and a 4.3% effect represents the impact for users (excluding potential outliers) with exposure to the experiment for a full 3-4 weeks over the experimental period. Less than two weeks in the experiment would have provided less than optimal exposure to the Goals feature, and if enrolling in Steady mid-week, the first week of exposure a user would not have received the goal-related push messages.

It didn’t matter which Goal variant was used. Both were statistically significantly higher than the control (p<0.015), with no difference between variants. Other key metrics (such as number of app opens, and number of job applications submitted through the app) were largely unaffected by the Goals feature.

In addition, the Goals feature appears to have a greater impact over time. The characteristics of the people in each condition were the same, and as people had time to set and track their goals, their income ultimately rose.

We included ~80,000 Steady users in the experiment, with ~40,000 in the control, and ~20,000 in each of the Goals variants. The bank account link rate was ~36%. We ran the experiment over 4 weeks, so over the course of that time frame ~14,400 users increased their incomes $28-$80 (2-6%) over that month, for an overall increase in income of $400,000-$1.15M due to the Goals feature.

Since the experiment, Steady has rolled the Goals feature out to its entire user base and through modifications to its onboarding has raised bank account linking to ~45%. We estimate that since August 2020, on average 36,000 users per month have increased their incomes $28-$80, for an additional increase in income of $1M-$2.88M per month due to the Goals feature.