Beacon

Discovering a new way to acquire data from 0 to 1

Overview

Delphia’s proprietary AI is used by our in-house investment team to predict the performance of stocks to give customers an edge over the stock market. The AI analyzes spend data to identify and predict purchasing trends before changes occur in the stock market.

Crowdsourcing this data from users was difficult and we currently had low traction. The Beacon team was formed as a 0 to 1 product team with the goal to find a new way to increase data contributions.

Product Design Lead

Beacon Team: Anastasia Artemova, Senior Product Manager & Nick Hallmark, Data Scientist

Duration: January to March 2023 (3 months)

Responsibilities: Discovery, user research, UX design, interaction design, information architecture, product vision and strategy.

Problems

  • Delphia’s investment portfolios were long-term portfolios which took about a year for an investor to experience value. This made it difficult to demonstrate immediate value for investment customers and to prove the benefits of sharing their individual spend data.

  • Experiments around gamifying data contributions by offering sweepstake tickets and crypto tokens also proved to be ineffective as they were only extrinsically motivating. Traction for these features were low, reduced trust from the experience investors, and drew in low-retention users who just wanted to try to make quick cash.

Business goals

Develop a new acquisition method that could increase data contributions.

User goals

Develop a new acquisition method that solved a need for our target users that was intrinsically motivating enough to encourage data contributions.

Navigating ambiguity

We were given no constraints for this except that the initial scope for the product would be an MVP and so, we began by setting our own constraints:

  • Connected to mission of building wealth with data

  • Target Americans initially as our existing data contribution infrastructure already allowed for U.S. customers

Identifying the “people problem” to solve

To align our new team, I conducted an initial brainstorm to prompt us to think about the “people problem” that could be solved by acquiring user spend data. We leveraged Julie Zhuo’s definition of “people problem” as something that many actual people are experiencing that our solution will address.

  • In this brainstorm, we considered financial, lifestyle, and professional problems to begin generating early hypotheses on problem spaces

  • The biggest themes to emerge were those regarding the current economic climate and financial stress

Considering the market and timing

We objectively looked at Delphia’s mission to help people build wealth with their data and considered the market and timing of this new product. During our secondary research, two questions began to emerge:

  • Is building wealth top-of-mind for a majority of Americans right now?

  • At what point in your financial journey do you need to be in to be in in order to start thinking about wealth building?

Secondary research to inform our hypotheses

As we conducted secondary research on what was top-of-mind for Americans when it came to finances, I developed a map that outlined the “Journey to Wealth”. This map contextualized our findings and was used to develop hypotheses and assumptions to test during our research sessions.

This exercise helped us identify that our current product offerings were tailored to a niche audience who were users that were later in their wealth journey. A problem this created was that the data collected from this audience may be skewed for this niche audience and that there was potential to reach wider audiences that felt more representative of America.

Hypotheses

  • We believe it’s challenging for our target user to manage their finances to reach their goals, especially during the current times of economic uncertainty.

  • The tools or methods available today are time-consuming and only provide information about what they’ve spent on after they’ve spent it. It’s requires a lot of effort to proactively be informed to make better financial decisions.

Before starting our research, we leveraged our secondary research findings to begin hypothesizing an initial target audience who would feel this problem the most.

Hypothesized target audience

  • 25 to 34 years old

  • Live in metropolitan cities with a high cost-of-living

  • High expense to income ratio (e.g. I make 100k but expenses leave me with only a few hundred excess)

  • Are highly social and desire the flexibility to enjoy their lives

  • Have “short-term” financial goals (e.g. paying off debts, saving more, or spending less)

  • Have an ambition for wealth and actively working towards building wealth
    Actively doing activities to optimize their saving and spending (e.g. budgeting, expense tracking, deal hunting)

  • Are financially and data-literate

Hypothesized factors which influence our target user’s decision-making

  • Financial literacy and money mindset

  • Apps and online usage

  • Social financial comparison

  • Inflation and cost-of-living

  • Mental health (anxiety, stress)

  • Social impact (sustainability, environmental)

We conducted 10 user interviews with our hypothesized target audience and gathered the following findings

Key findings

  • Most participants used apps Mint but eventually went back to spreadsheets due to poor categorization

  • Expense tracking is used to identify behaviours they want to change or funds they want to re-allocate in the future

  • Economy is a major stressor. Inflation has impacted the way people spend.

  • Participants go through a cost-benefit-analysis before making large purchases, however they feel more “trigger happy” for small purchases

  • Spending habits are not top-of-mind

  • There’s a desire to learn from others who have come from similar situations – it makes the advice feel more tangible and “real”

Initial “ideal” target audience

  • High spend to income ratio

  • Finances are top-of-mind. Have goals they want to achieve, concerned about economic situation.

  • Needs to be highly motivated to want to make a change in their lives

  • Actively works on self-improvement

  • Track parts of their lives already (e.g. fitness, sleep)

  • Actively up-skills both in career and life

  • Desire to spend consciously

A pattern that began to emerge was the pattern of spending and saving for our target audience. We noticed a behaviour where our participants were manually tracking their spend to identify behaviours in their spending habits. This raised the question for our team: Could we do this work for them?

Making sense of spending data

Our Data Scientist, Nick, developed our first proof of concept and correlated his spending data, emails, web browser history, and fitness app data to identify behavioural insights. Nick was already tracking information about many of his lifestyle habits for his fitness tracking app and used this to test our hypothesis that there was a correlation between external behaviours and spend.

This exercised also helped expose potential insights we could derive by correlating behavioural, spend, and digital data. Some of the most interesting datapoints included:

  • How working overtime, sleep, and alcohol consumption impacted take-out and grocery spend

  • How frequency of website visits impacted purchases probability of purchases

How this connects back to AI investing

One of the constraints we gave ourselves was to connect our new product solution back to our goal of building wealth. As a gut-check, we wanted to ensure our solution laddered up to our business goals and so we mapped out how the datapoints gathered for this behavioural spending app could map back to the data that was most beneficial for our AI to use.

Laying the foundation for the user experience

Taking all of our learnings from research, I began to explore what this would look like in an app experience and considered the following factors while designing this flow.

  • How might we build immediate trust and value to our users?

  • How might we make this experience engaging and habitual?

  • How might we do the heavy lifting for users when it comes to expense tracking and budgeting?