[TIL] Turning TodayPoor toward shared spending

What I did today

• Check feedback on the first plan

• Discussion of direction for planning and supplementation

Key feedback received

• The most common part of the feedback was that the direction of this service needs to be clarified.

  • Is it a service to improve consumption habits?

  • Is it a fun SNS that shares consumption history?

• After discussion, our team decided that it would be right to move closer to a fun SNS that shares consumption history.
  • Improving consumption is an additional effect, but the key is to share and compare consumption with friends and make it fun to keep records.

• OCR itself is already provided by many household account book services or receipt record services.

• Therefore, what differentiates TodayPoor is not simply recognizing consumption details through OCR;

  • Friends-based consumption sharing

  • Random consumption topics

  • AI roasting/upward feedback

  • MVP selection

  • Continuity of records through social pressure and fun

  We rearrange the point that it is in

• In other words, OCR is a tool that lowers the entry barrier, and the core of the service can be seen as the experience of sharing consumption like a meme

• The initial plan included many features such as OCR, SNS feed, likes, comments, AI MVP selection, AI feedback, random topics, chat, etc.
  • However, we received feedback that the functionality may be too broad considering the 6-week project period.
• So we decided to focus on the next flow in MVP.

  1. Upload consumption history capture

  2. OCR extraction and user modification

  3. Create a daily consumption feed

  4. Random topic participation

  5. AI roasting/consolation feedback and MVP selection within the group

  Advanced SNS features such as chatting, likes, and comments will be given a lower priority.

• Because consumption details are sensitive information, there was also feedback that designing the scope of disclosure is important.

• In particular, disclosing all information such as amount, affiliated store, category, etc. can be burdensome, so users should be able to choose whether to disclose it or not.

• Even if it is a fun service, I feel that it is important to have safety measures in place to prevent users from feeling shame or discomfort.

• MVP should at least consider the following features:

  • Choose whether to disclose information

  • Whether to agree to AI roasting

  • Good AI/Bad AI mode selection### What I learned about ERD

• Currently, AI functions have two main roles, but in ERD it was not clear where personal feedback connects.

  1. AI results for MVP selection

  2. AI feedback on personal consumption history

• Therefore, if personal feedback is included in the scope of MVP, EXPENSE or POST and AI_RESULT must be connected, and if AI feedback is attached only to MVP results, it is sufficient to connect to RANKING_RESULT as is currently done.

• We also decided that it would be a good idea to add the mode column to AI_RESULT to distinguish between good AI and bad AI modes.

Team discussion results

• Based on feedback, we defined TodayPoor as closer to a consumption-based social game/SNS than a consumption habit improvement app.

• In addition, compared to similar services, it is concluded that the structure of sharing consumption with friends and AI providing interesting feedback is a more important differentiator than the OCR function itself.

• In the future, the following information must be supplemented in the plan:

  • Clarification of service purpose

  • Persona specification

  • KPI settings

  • Addition of personal information disclosure scope policy

  • AI_RESULT structure supplementation

  • Reorganized MVP feature scope

Lessons learned

• Through this feedback, I felt that it was not enough to just have a fun idea, and that the purpose and core experience of the service had to be clear.

• At first, it seemed like a good idea to include a lot of OCR, AI, and SNS features, but in reality, we had to be able to explain why users should continue to use this service.

• The point of TodayPoor is not to accurately analyze consumption, but to have friends look back on their consumption in a funny way and keep recording it.

• When supplementing the plan in the future, rather than adding a lot of functions, I will organize it so that the core user experience is clearly visible.

What to do next

• Meetings by series

  • Modify ERD by reflecting feedback

  • Writing API specifications