MVP has too many mandatory features + Technical Hurdles?

[fullofglory]

An MVP I'm building is an E-commerce platform, but...anyone who knows an E-commerce platform is not an easy feat and actually some people's main offering. E-commerce is just single module of what i'm trying to do. The main goal is an entire ERP system.

I have a full-time job as a frontend dev which is nice as knowledge from job can be used to build my startup.

My stack and skills (Skill Level Rated as 5 being best):

1. 3.5/5 - UI/UX: Any design tool, but using a TailwindCSS library (made mistake of trying to build my own design system earlier, massive time waster)
2. 4.7/5 - Frontend: HTML5, CSS3, TailwindCSS, Typescript, React, Next.js
3. 4/5 - Backend: GraphQL, Node.js, Fastify, Typescript, Prisma
4. 0/5 - 3rd party integrations: Stripe, Paypal, some sort of maps API, didn't decide yet.
5. 2.7/5 - Data Modeling - Postgres
6. 3.5/5 - DevOps: Docker, Github Actions, Terraform
7. 2/5 - Cloud Infrastructure: AWS
8. 1/5 - Monitoring: Prometheus, Loki, Grafana

Target Market:

1. Me as customer archetype and customers from job, massive need for this and nearly applies to every business. Potential is huge.

Questions:

1. What do I do if my MVP contains many mandatory features just to get it functioning?
2. Big technical hurdles, how to circumvent? As you can see, my main skill set is the frontend and drops off the deeper you start going into operations. What do I do, if I have big technical hurdles?
   1. Data modeling business logic - probably the hardest thing to do and something I struggle with. Not sure if I'm struggling because my data model is pretty much enterprise level. I have like 400+ tables in Postgres.
   2. Infrastructure - I lack experience here just because, playing with infrastructure is where the costs begin. My architecture is mainly serverless, but some things can't be serverless like ElastiCache or RDS. I know there is Aurora serverless for relational databases (but I heard the costs are signficantly more expensive than RDS. Thats why I'm opting for RDS multi-AZ instead. Due to the serverless nature, I feel like I am forced to do microservices, but anyone with technical knowledge knows, microservices solves a bunch of issues, but also brings a whole host of new issues and dramatically slows you down. If I don't go microservice with serverless, I'm essentially having an anti-pattern (Monolith Lambda Docker). I want services that scale to zero if possible.
      1. So idk, monolith or microservice?
   3. Monitoring - Don't have monitoring experience because haven't deployed my app yet. Only know that I want to most likely do Prometheus + Loki + Grafana

 

[patriciadsykes79]

@fullofglory If the MVP you've come up with contains too many mandatory features, you're thinking way too big. You can definitely make it smaller.

You need to find out the absolute must have feature(s) that your initial customers will want to solve actual problems people will pay/use you for.

I would suggest going back a step and thinking about the problem you're trying to solve for those initial customers. If it's just another general e-commerce platform, I think you could really struggle (even if it is just intending to be a module, the next steps will only happen if this bit is successful).

Also in the 'infrastructure' section, this all sounds far too complex for an MVP. Try using a PaaS (platform-as-a-service) like Vercel, Netlify, etc to get started before investing time in anything like this.

 

[fullofglory]

@patriciadsykes79 A lot of these are must-haves, like can't have an E-commerce platform without payments, am I right? And a lot of the features are USP, meaning they are what is going to differentiate me from he masses. This stops it from being a general E-commerce platform. E-commerce actually is only one part of the app. Like I said the main goal is ERP, but thats huge. Thus, E-commerce module being the MVP.

 

[patriciadsykes79]

@fullofglory Well yeah I would imagine payments would be required, but I'm sure a fair few of the other things on the list could go, at least initially.

I would personally pick one or two of these USPs maximum and find a way to make that work as an MVP. i.e. digital sales only, cutting down on the need to build shipping, address stuff, etc...

The decision is ultimately yours (you know your target customers), but if it's going to take too long to build, you risk not being able to make it sell after all of that effort.

 

[fullofglory]

@patriciadsykes79 Forgot to answer infrastructure.

Yeah my fear for not using something simple like Vercel is that let’s say my app scales. Vercel’s pricing scales exponentially. That’s how they get yah. So I didn’t want to not know infrastructure and go with Vercel and have the price exponentially grow and I’m like…crap I don’t know how migrate off of this. Don’t want to be scrambling. Also harder to integrate with AWS, I wanted to use AWS WAF for layer 7 Ddos protection. But obviously Vercel doesn’t give you that type of control.

I guess I could use Vercel to deploy frontend, and have it talk to my AWS API, still doesn’t solve backend deployment complexity.

 

[patriciadsykes79]

@fullofglory Have you seen https://www.flightcontrol.dev/? It might help you out with that infrastructure issue with the ease of PaaS!

 

[ericbunic]

@fullofglory Hi there - cofounder of https://www.withcoherence.com/ here. Similar to what you've posted above but we support both GCP and AWS along with cloud development environments which make it easy to write, test and debug code from the same dashboard. Would appreciate any feedback

 

[walkingwiththeholyspirit]

@fullofglory

1. Some MVPs are large, minimum is relative, so it’s normal don’t worry

For instance if tomorrow I wanted to launch a CRM with a unique positioning, I’ll still have to build a core expected feature set that is enormous, otherwise it wouldn’t be an MVP, just a prototype

The size of the MVP will mainly depends on:
- what’s the minimum solutions to correctly solve the problem
- what’s the minimum to be competitive

1. 400 tables ? That’s kind of crazy
   Unfortunately I’m in the same position as you (frontend dev trying to learn how to do back / DB / Cloud for production), and the best you can do is learn from experts

You should outreach people knowing those subjects very well and ask for help in your specific situation

If done right, people will give you excellent advices

I also heard that AWS have sales that you can connect with, and they connect you with solution architect helping you build your cloud infrastructure for free, might be something great to try !

1. I can’t help you on this one, I’m also trying to understand how to setup those correctly

 

[613jono]

@fullofglory an MVP is not an APP