The new gold rush: go from zero to startup in a couple of hours and for less than $100.

In this article, I share all of the ingredients needed to research, design, build and launch your startup idea.

True confession: I just backspaced out the sentence “Back in my day it would have cost $20k to build a working software prototype.” Saying that made me feel old. But it is true. Not just that I’m old. But that the world has profoundly changed for the better. We have opportunities to build technology-focused apps that no one has ever had before. The barriers to entry have never been lower. It is the gold rush all over again.

Just five or so years ago, if we wanted to put a real product in front of a real customer and be in a position to ask them for real money, we’d have to spend a few hours coding (and potentially learning to code) or hire a couple of programmers to build something out. We’d also have to host this app and do all of the wiring to get it live. Oh and we’d be basing this on a product strategy built on a couple google sessions and gut feeling.

Now? We can research fast. And design fast. And built fast. And launch fast.

This past Tuesday, I was privileged to host a session at Bentley University on this topic. To market the event, I promised we’d take a student idea and build a product strategy around it. Then, we’d build a working prototype we could put in front of interested customers. That was a big promise that, in full disclosure, I wasn’t 100% confident I would fulfill. But, lo and behold, it worked! We built a decent prototype that would - with some backend changes - fulfill the problem we set out to solve. Here’s how we did it and the tools we used. Of course, you should receive this plan as merely my thinking and let it inspire your approach.

Tools:

• LLM. I used Claude as a strategic thought partner. I like Claude although at times it can be too verbose and often lean hard into my ideas rather than tell me an idea is a bad one. I use chatGPT a lot too, and find it to be helpful. Either would likely work for this exercise.

• “Vibe-coding”. I used Loveable to build the prototype. I think it does a good job interpreting ambiguous product requests and properly building modern app experiences. I use it for other projects and while credits can add up fast, it’s possible to use it. For $60, you could use it for two months with more credits and then downgrade to make minor changes in incremental fashion.

Pre-Requisites:

I follow the following steps to shape the idea.

Important pre-requisite: Each step depends on pre-loaded (or attached) instructions that are at the bottom of this article. Load the instructions first.

Keep in mind a few things:

1. I favor simplicity. The frameworks, definitions, estimates, outcomes are all purposely simple.

2. You will likely need/want to re-prompt, go deeper into a topic or ask Claude to adjust based on a business goal. For instance, in the example we did in our session, our idea failed to reach profitability. Realistically, I wouldn’t have built a prototype before figuring out a sound business plan. This would require a few repetitions with your LLM.

3. If you get stuck and can’t get clarity around a good business/product strategy, consider that a good thing. Maybe you need to rethink your idea, your model, your market or something else. Better to learn this now, then after you’ve built a prototype.

4. At every step, when you make selections about the business (the strategy, the customer, the positioning statement), you must tell your LLM so it knows moving forward.

5. Have fun with this and make sure you adjust your instructions to your liking.

Building a Strategy with an LLM:

1. Open your LLM of choice and load the instructions at the end of this document (edited to your liking, of course).

2. Write out your business idea in plain text. This is the basis for all the remaining steps as we transform this idea into a working project. Use as much detail as you’d like and try to identify the problems you’re solving, how your solution might work and who would stand to benefit from your idea.

3. Construct your problem statement. Turn your business idea into a problem statement. Paste your business idea into the LLM and ask the LLM to create a problem statement based on it.

4. Perform an opportunity analysis. With your problem statement in hand, ask the LLM to list the opportunities. It’s possible you’ve only got one. But likely, there’s a few things you could do with the idea and it’s important to list those out so we have somewhere to start. The instructions (which you’ve pre-loaded) include a weighted scorecard that will priority rank the opportunities you can pursue. Tell the LLM which opportunity you’d like to pursue with your business.

5. Build a competitive analysis. Now we need to know who solves that problem today. Not just potential competitors but other substitute solutions (public transit as a “competitor” to ride-sharing for example. We also want to know if the customer solves this on their own. The competitive analysis will draw this out and - most important - start to help you figure out why you’re different.

6. Pick a strategy. Now that we know the competitive space, we can figure out our corporate strategy. Will we just be less expensive? Unique in a new way? Focused on the broad population? Or just a sliver of the market? Prompt the LLM to help you devise this piece.

7. Choose a differentiator. This step is closely coupled with the one before it. If we’re differentiated, then how? With our competitive analysis we can begin to lay out opportunities to differentiate. Note: if none exist, maybe we circle back to step 4 and pick a different opportunity.

8. Define your customer. Of course, we need to sell/cater to someone. Let’s figure out who our customer is. Build out an ICP (Ideal Customer Profile) with the LLM. An ICP is a definition of your customer that is likely purchasing your product. You could have a primary user persona too, and sometimes it’s the same person. With the market, problem, opportunity and differentiator defined, it might be rather easy for the LLM to pick out the person who’d be interested in your company. In the marketing world, an ICP can be a lengthy, detailed description of the customer. As promised, our instructions keep this relatively simple.

9. Build a positioning statement. A positioning statement is your reason for being. It should tell the market - and prospective customers - why you exist and why they should choose you. A lot of marketing material will dangle from this asset so make sure you like it. Once again, ask the LLM to build you a few positioning statements and choose one you like.

10. Figure out financial feasibility. Now it’s important to know if your company can make money. Ask your LLM to run a feasibility study. It will report back estimated income, expenses and - of course - profit. And it will do this over the next few years. If you’d prefer to have positive income sooner, work with your LLM to drive costs down and revenue up. This step, more so than the others, may take a few rounds of prompting. When you have a feasibility study that is a) realistic and b) matches your financial goals, it’s safe to move on.

11. Plan your product roadmap. Your product roadmap is a high level plan of when you’ll build features/capabilities/additional products over time. This roadmap should correlate nicely with your feasibility plan, especially if you plan on price increases, new product offering or expanding into new markets. The product you start with won’t magically scale with these changes.

12. Focus on an MVP. An “MVP” stands for your “minimal viable product”. Essentially, it’s a definition of the smallest version of your product that would still be valuable to a paying customer. Building an MVP requires discipline and focus to ensure that your product only includes enough code to satisfy the customer problem you defined at the start. This is an important artifact as it is what you will feed to your vibe coding tool to build out the prototype. Once again, you may want to redefine your MVP as you get feedback from prospective customers.

13. Craft a marketing pitch. When we’re ready to launch, we’ll want to sweeten the deal for people coming to the site. There are a lot of assets we can build out to talk about our product effectively. But one compelling piece is an ROI stat. If possible, ask the LLM if it can calculate the potential ROI a customer would enjoy if s/he purchased your tool. You’ll want to word this carefully if the data isn’t verified. However the financial upside of your solution should far transcend the cost of the software you’re building. It’s important to educate your audience as to what that is.

Create naming and marketing assets with an LLM:

1. Build a brand. Of course we need to name this thing of ours. Use a tool like chatGPT to brainstorm some names. Or - search for available domains to land a good name. Ideally your name will be easy to remember, spell out and memorable for your customers. It’s also helpful if your name describes the solution you provide. You may also want to create a quick logo to dress up the site and for social media. ChatGPT can do this quite easily. Take some of the assets from the previous steps, mixed with your name and do a few rounds to get a logo that works. Lastly, secure a domain for your site. If you can’t get your exact name, find one with the word “try” before your name or look for a domain extension outside of the popular .com’s. For the first year, a domain shouldn’t cost too much, maybe $10 if it’s an available domain.

Build a working app with vibe-coding tools:

Finally, we have enough to build the first iteration of our app. While building business strategy is fun (for me, anyway), building working apps in a few minutes is equally exciting! Identify a vibe-coding app. As noted earlier, I prefer Loveable at the moment. Some will cost money to use fully, but look for something that fits your budget.

1. Start a new project. You have an option right off the bat: Do you want to create a working app or a marketing page to tease your product? You could combine your working product with the marketing home page into one app experience if your app is small enough. Including a “try the product” button allows the user to go from the marketing page to the app. Alternatively, you could start two projects, one for a marketing page and one for the app. This may get complicated for a prototype so I’d lean toward a combined offering or just building the application for the purpose of validating it with customers. Name the project whatever you like. With Loveable, they’ll start with some initial prompts.

2. Start prompting. Here’s where we want to describe the problem you solve, the ICP and your MVP definition. Ask your vibe-coding tool to build a solution using your problem statement, MVP definition and ICP. Hit go and let the vibe coding begin. Your tool will likely take a few minutes to build the initial experience but rest assured that your vision is starting to come together without having written one line of code.

3. Iterate. Once you get back a working app to play with, work through the app. Does it function the way you had hoped? Is anything missing? Are there small bugs or nuances? Get comfortable working with your vibe coding platform and providing prompts to modify your app. Be wary of credits however as these can run out fast. Try and keep focus on optimizing the MVP experience and building a tool that can satisfy the ROI promise built out in a previous step.

4. Consider other features. Depending on how/if you’re monetizing your product, you may want to introduce other features. For example:

   1. If you wish to have users log in on their own to use the product (as opposed to you demoing it to them), Loveable can easily implement basic (but secure) authentication.

   2. If you want to test “willingness to pay” and ask customers to pay for your service, you can easily integrate a Stripe account and payment process into your experience.

   3. Maybe you want to measure how users are using your product? Telemetry is important to understand usage/adoption/attrition. I haven’t done this one myself but it’s on my list to get to soon.

You should be good to go now with a working product. Best of luck and have fun evolving your idea into a profound business. One last reminder: As of today, vibe coding tools will only get you so far. As you scale and seek other services, you may need to build something more robust. Hopefully the steps above can help you get started, validate your idea and move you to the next stage of the journey. Good luck and enjoy!

Instructions to load into your LLM:

If using Claude, create a project and copy the text below into the Instructions section for a project. If using chatGPT, create a project and upload the instructions below as a file named “Instructions”.

Problem Statement

Build a problem statement using the following format: “[Who] struggles with [what] in [when], leading to [result], because [why].”

Opportunity Analysis

Score each problem using these weights:

• Current demand to solve the issue: 40%

• Cost to implement solution and enter market: 30%

• Competitive intensity: 30%

Competitive Analysis

Map each competitor and substitute with:

• Core strategy (using Michael Porter’s generic strategies framework)

• Value proposition

• Strengths and weaknesses

Include competitors, substitutes, and DIY methods.

Corporate Strategy

Use Michael Porter’s competitive strategy framework (competitive advantage and competitive scope) to select one quadrant:

• Narrow or Broad scope

• Cost leadership or Differentiation

Differentiator Selection

Choose from: price, features (specify which), quality (specify: reliable, fast, secure, accurate, etc.), experience, service, brand, innovation, values, community.

Ideal Customer Profile (ICP)

Define:

• Job title

• Job description

• Core problems

• Success metrics

• Constraints

Positioning Statement

Use the following format: “For [ICP] who struggle with [Problem Statement], [Product] is a [Category] that [Outcome]. Unlike [Competitor], we [Key Differentiator].”

Feasibility Analysis

Create a table showing across five years:

• Overall costs (operational, product development, COGS, marketing/sales, support)

• Overall projected revenue

• Number of customers

• Overall profit

Include detailed breakout of each item below the table.

Product Roadmap

Create a table with:

• Capability themes (rows)

• Time horizons: Now, Next, Later (columns)

Market Analysis

Define and calculate:

• TAM (Total Addressable Market): Total demand if capturing 100% of all possible customers

• SAM (Serviceable Available Market): Portion realistically servable given product, business model, and geography

• SOM (Serviceable Obtainable Market): Portion realistically capturable in near to mid term, considering competition and execution

• STM (Serviceable Testable Market): Hyper-targeted market for solution validation

MVP Specifications

Write specs for the smallest version that:

• Provides value

• Meets value expectations of ICP

• Aligns with starting price point from feasibility analysis

Include only necessary functionality as identified in the opportunity analysis phase.

ROI Analysis

Create a simple ROI calculation that:

• Aligns with ICP’s success metrics

• Uses returns generated from MVP solution

• Can be advertised to prospective ICPs

Calculate using: (Return - Investment) / Investment