Welcome to the first blog into my series where I go through all the stages of hardware product development - from that persistent idea at the back of your mind to the triumphant launch of your product. Join me as we navigate the highs and lows of bringing a dream to life.
Idea
It's that idea again. The one you try to shake off, but it just keeps coming back, more persistent each time. You realize you need this solution for yourself, and it's glaringly absent from the market. So, you think, "Fine, I'll build it myself". But how to start? You turn to ChatGPT and utilize this powerful tool. "Here's my idea," you start, sharing your vision and providing as much context as possible to get a better response. "I can't stop thinking about it. How do I bring it to life? Can you give me a roadmap, a step-by-step guide?" After reading the advice and asking more questions, you find yourself armed with a solid understanding—a clear roadmap but still theoretical. Now, it's time to take action.
My journey began with a dead phone while I was on the go. I couldn't find a solution to charge it, and I wasn't alone. So...I decided to build a phone charging solution for public places.
Validate Your Idea
Before diving deep validate your idea. Ask yourself: What problem does my solution address? Who am I creating this for? The most impactful products are those we build to solve our own problems. Begin by sharing your concept with close contacts facing similar problems. Gradually extend your reach, moving from familiar circles to the "scary" task of engaging with strangers. Y Combinator highlights, the primary reason startups fail is their reluctance to engage with customers.
Now, it's about selling. But what are you selling? Essentially, a prototype as basic as possible - crafted from nothing but duct tape and sheer will - think simple and quick, using whatever materials you have. Spending more than a week on your initial prototype is a miss. Temporary brokenness > Permanent paralysis. Perfectionism is fear of criticism. Aim to create a basic version of your product within a week, and in the following week, pitch it to your identified target audience. The objective is rapid feedback from target user.
Focus on the users' interaction with your product rather than their words. Comments like, "It's cool, but I'd buy it if it had this feature." - red flag. If you get many of them pain doesn’t exist or maybe the wrong crowd to sell to. Does your product solve real problem real need? Did you correctly identified target customer? Once you believe your idea solves a real pain, it's time to progress to the next phase.
No matter what is the outcome congratulations you pass first test you talked with customers!
I started with a wireless charger I bought on Alibaba. I personally installed it in the first café that agreed to test it, learned a ton, realized the problem existed, but wireless charging was not the right solution—so, I moved on to portable battery chargers provided by fully automated stations that do not require any management from location employees.
Go-to-market strategy
You've stepped into the arena, built an MVP, and talked with your first customers. You've learned a lot and validated your idea. Now, it's time to create a grand plan to conquer the world. Clear your calendar and, together with Google and AI tools like ChatGBT, create a go-to-market plan.
Who are you selling to? Who's your target customer persona? How much do you want to sell it for? What's your distribution strategy (direct to consumer, or through channels like retail, etc.)? What's your marketing strategy? How can you sell the first batch before you've even produced it? How are you going to fund R&D? This is a basic business plan just to understand how it works. I would spend a full week on this strategy and then revisit it for an hour each week to update it based on what I've learned.
Beware of analysis paralysis. You will learn at every step, and your trajectory will change. I started with the pay-per-use model, where users, for example, pay $1 for 24-hour access to the chargers, and ended with a hardware-as-a-service model where locations pay us a monthly fee and provide chargers to their visitors free of charge. Starting with the wrong hypothesis in the beginning and learning is better than not starting at all.
Design and Development
Now the fun part begins. But not to get carried away start with understanding restraints such as cost, regulations, processes, manufacturing, logistics.
What environment will your product operate in?
You must be very clear about the conditions your product will operate in: temperature extremes, humidity levels, dust and water exposure, shock and vibration resistance, electromagnetic Interference (EMI), altitude effects, chemical exposures.
How are you going to build the team?
I knew I couldn’t afford passengers. I knew exactly what skills we needed to build it. I wanted to build the smallest team possible, where each person knows their area of expertise from top to bottom. A-players. My goal was to build a special forces team with whom I go into unknown territory where I don't have time to hold hands. Each person manages him or herself. How to get those people? They are ambitious, have their goals, and they look for adventure. If your goal is ambitious and achieving it helps them to achieve their goal, you are aligned. The next step: do they believe you have what it takes to lead them? Do you have something that they don’t that makes you not just another dreamer? Something that will help these high performers to weigh all the factors, risk their time, and join the journey?
Who should be in your team?
To build a consumer electronics product, you will need an industrial designer, electrical engineer, software and firmware engineer (perfect if this person is 2 in one), if you have some special functionality that requires an expert, you need that, marketing/sales which better be one of the founders, manufacturing outsourced.
S.Jobs (vision/engine), S.Wozniak (genius all-in-one engineer),
M.Markkula (money/experience)
How to avoid scope creep?
Scope creep refers to the gradual expansion of the project scope beyond its original objectives, often resulting in increased costs, delayed timelines, and diluted project goals. One of the most important and hard tasks is the ability to say NO to a lot of great ideas. Your job is to get your product to customers who need the core functionality as soon as possible. Anything that is not the core ruthlessly remove to a backlog of ideas. Remember, there is always version 2 and 3 and 4 etc. where you can add all bells and whistles. The first iPhone couldn’t make videos. Now I made all my videos about "How I built Supernova" on iPhone 12…I made this mistake with Supernova with a couple of nice-to-have features that some took more than must-have features…Don't be me.
Identify what is your killer feature and spend most of your energy on that. You are not Apple; you don't have decades of expertise, unlimited resources, and a world-class team and supply chain. Lucky if you have one of these. And also, every new feature requires deep testing.
Place some restraint from the beginning; they will force you to be creative and not bloat the scope. One of those is the target price you have from your go-to-market strategy. Since you have the price that you think users will buy it for, it will limit your appetite to add more features which equals more components, more operations, and higher price BOM and final cost.
How to make hard technical decisions?
You will be faced with a few very hard-to-reverse architectural decisions. Like, do we use WiFi or Cellular for connectivity? What’s the power source, a wall plug, one-use, or rechargeable batteries? Does software that controls the device run locally or in the cloud? Decisions like this will have 3-6 month development consequences and also certification if the decision later reversed must be done again. First, you will need to identify those types of decisions. Second, put a deadline for this decision to avoid analysis paralysis. Third, within the timeframe, do a deep 360 research involving all sides to gather as much information as possible. Analyze this information and gather the opinion of every team member. Finally, make an informed decision. And once you've made it, fully focus on execution. It is not guaranteed you will be right, but the goal is to get the best possible decision within a reasonable timeframe. If you did everything you could but the market thought you were wrong, that’s totally ok.
The first version of our Electron charger was based on Bluetooth - each battery activated via Bluetooth by the app. Then we learned people won’t download the app to charge their phone and chargers must be locked because no one has time to manage the chargers. We made the station Supernova WiFi-based with a lock. Then we learned that it is extremely difficult to get WiFi from locations and installations/onboarding of new locations is way too complicated. So we redesigned and added Cellular and had to certify FCC twice, first for a WiFi device and then for cellular. We made the best possible decision with the information we had. If we didn’t launch, we would never have learned it.
How to make the transition from prototype to mass production as fast as possible?
To build faster, you must involve the factory as soon as possible. It would be a shame if you made a prototype, sold it, and then, when it comes to manufacturing, you learned from factories that you need to redesign the entire product to make it manufacturable for your target price. This happens all the time. It's easy to build a single item of anything. It is magnitudes harder to build thousands of something. A totally different level of complexity. The cost of every decision and mistake is multiplied by the number of units you plan to produce. I cannot stress enough how important it is to involve the manufacturer in the process as soon as possible unless you are Apple and can invent a new manufacturing process just for your product.
Off-the-shelf vs. Fabricated
As fewer customized components you have, the easier it is to manufacture. Each customized fabricated part requires design, development, performance test, design manufacturing process, design quality control process. It's infinitely harder than buying off-the-shelf components. If off-the-shelf components are available, always choose off-the-shelf. In general, use customized only for core functionality. Your goal is to have a product like Lego made out of ready already battle-hardened parts.
How to produce a high-quality product?
"The best code is the code we didn't write" - Steve Jobs. The moral of the story is to keep it as simple as possible. Use off-the-shelf components. Build with an experienced factory. Trial run production and quality control.
Regulation and Certification
Keep in mind there is certification involved. If you have any radio, you will need to certify it. If you have a power supply it must be certified. If you have a battery, shipping will be tough and require certified packaging and marking. Certifications such as Apple's Made for iPhone (MFi) are enforced if you want to create products using their proprietary technology. Additionally, some retailers might request additional certifications - if you want to work with them, research what they require as soon as possible.
All components that require certification must be bought off the shelf unless it's a core functionality. If you are building a new power supply, fair enough, need to certify, but if the power supply is just a part, find off-the-shelf and optimize for the existing part. We were thinking of make special power supply for our station - supplier told it will require certification. I asked how much? Estimated cost 15000-25000 American dollars...No, thank you!
Regulators require to have a certain marking on the product. You will need to know what type of marking, dimensions and choose the area on the product and package to print it following the regulations. So you must confirm it before you produce.
Manual
Creating a good manual is extremely important and not a simple task. A good manual will save your support team tons of time. Start the draft as early as possible. It's a good exercise that will help you clarify the scope and see if the product makes sense because explaining how to use it in the manual will reveal things you haven't thought about. Once the product and consequently the draft are finalized, find someone experienced in manual making and hire someone with experience on freelance platforms. We found a great guy who did manuals for companies like Canon on Upwork and leveraged their expertise.
Packaging
This is overlooked but is very important. Make packaging that is not only good-looking but is standard to not exceed standard pallet rates; that's what happened to us. If your package intends to be used many times like ours, you need to design a reusable package. If you have a battery in the product, you will need a specially certified package with special marking. In our case, it was a nightmare we had to divide the package because batteries cost more to ship.
Launch
The goal is to collect preorders - the goal is ideally not to have a warehouse - sell before you produce. It’s assumed that product launch looks like the rocket launch with thousands of people glued to the screens anticipating the event. It can’t be further from the truth. Apple and Steve Jobs specifically distorted perception with their bigger than the world presentations of the new product. But as a new company, you should follow what they did for their first product, Apple I and Apple II. It was presented multiple times. You will have multiple launches ideally as frequently as possible. The launch date is a good milestone as a deadline though for all the team to sharpen the focus. But you must be launching and presenting all the time. Assuming you didn’t neglect marketing and sales during the design and development stage, you shouldn’t have a problem with launch.
There are different ways you can launch. But again the goal is to sell all your products before you launch. Even if you do crowdfunding through platforms like Kickstarter, you must start collecting preorders 6 months prior to launch. What you see as an overnight success was warmed up for years. One of the common strategies is to launch during shows and exhibitions. I did that with Electron multiple times by launching on biggest consumer electronics show in Hong Kong where I met one of my founding partners and than on CES in Las Vegas and etc. I presented Electron in one way or another on 50+ shows. The struggle is real.
A recent good example to study successful product launch is Rabbit. But I would recommend to study failed projects as well. I will talk about it more in my next posts.
Manufacturing
Every decision here will cost tens of thousands of dollars. When you place an order for even the smallest component, you buy thousands of them.
OK so you've collected preorders; now it's time to produce. Assuming you did the right thing and designed with the factory, your product is ready for mass production without hidden surprises. You've prepared your supply chain and know who produces what and which supplier is for which component. And if you kept a healthy ratio of fabricated to off-the-shelf parts, you should not be bogged down in details on manufacturing.
How to Choose a Manufacturing Partner?
This is one of the most important long-term partners. It's not easy to change. Don't chase big brands. Foxconn has a long list of A-level brands with stable order flow that they will prioritize over a small startup. The wiser decision at the beginning is to choose flexibility over other attributes because you will most likely be changing designs, volume is unknown, challenges are unknown, and you need partners who will commit resources to assist you in solving them together.
Choose a factory that has prior experience with products similar to what you are building. Leverage their experience and network of suppliers as much as possible.
How to Build a Relationship with Your Manufacturer?
Have clear requirements, be transparent, and communicate frequently. They can’t read your mind - be very clear about what you want. No one likes changes on the go, especially when juggling thousands of factory employees and hundreds of assembly lines, and you try to get the highest ROI from equipment. Whenever you make a change, it has ripple effects and sometimes diminishes the factory's profits. The further in the process, the more consequential. They will charge you to reload the assembly line. That’s why it is so important to involve factory personnel in the design process. They will not hesitate to tell you, “This is a good idea but impossible to produce for your target cost,” or “How are you going to test quality for this feature? We’ve never done it before, and we don’t know who has.” Make the factory part of the team and listen to them. You are competing for their attention as well. You are one of many customers - if they feel you are draining their resources that they’d rather give to 20% of customers that bring them 80% of revenue, you will be pushed down the priority list very quickly. This is how you build long-term relationships and get good results. This also reveals whether it is a good partner. Transparency reveals deficiencies on both sides. Be easy to work with.
As for the contract with the factory, I did not spend much time on it because I didn't have time. We had ongoing discussions, though, while we were working. One of the reasons I wasn't pushing to sign the contract is because I did not want to be tied down by it if the factory was not a good fit.
How to Identify if a Factory is a Good Fit?
Involve them asap to see how it goes in the development stage. If it feels good, proceed to the next step: run a trial, produce 10, then 100, and only then 1,000. This approach not only reveals issues at each stage before you produce 1,000 products with bugs or issues, but it also helps you see if it is a good factory fit. If it's not, be ruthless. Quickly prepare an alternative and move fast. I changed factories 4 times. In the most crucial, dangerous situations right before production, I was bold enough to just move. It looked scary and crazy from the outside, but it was the right decision, and we succeeded in the end.
How to create a quality control and production process.?
Trial and error. Start with production of 10 units of your product to find the bottlenecks in your process, to test the reliability of your design and components. Adjust based on what you learned before you place an order for thousands of components. Next, produce 100 units. The goal is to perfect assembly processes, identify issues, test components before you assemble them, and perfect the quality control process. Now that you did 100 and identified all the issues and conducted quality control, it's time to go ahead and produce 1,000. If you did a thorough job at the design level, sourcing, designing quality control, and assembly, things should go relatively smoothly.
For example, I identified that the glue our factory used was inadequate because we never paid attention to it, and glued components in the finished product were falling out. If we did not identify this issue with tens of assembled products, we would be in BIG trouble if we produced thousands...all because of one component - low quality glue.
How to Choose Suppliers?
Ideally, have 2 suppliers for each component in case something goes wrong with one. But that’s impossible in the beginning. You take the risk and go ahead. Thats why it is so important gradually increase production volume to catch the bad components as early as possible.
How Many Units to Produce?
Factories have minimal order quantity (MOQ), so you can't just order what you want; they need to make money.
Post Launch
Time to deliver sold products!
How to ship from factory floor to customer?
You need to pick a shipping company to deliver from your factory, which is most likely in China, to the country of your destination. The cheapest and slowest way is by sea, taking between 25 and 45 days to America or Europe.
The fastest and most expensive is by air. We used air delivery for time-sensitive deliveries. You'd better plan ahead and send everything by sea, because sending by air will dramatically increase your costs. The last mile in the country of destination is done by air or by land, land being the cheapest.
Third-Party Logistics (3PL) or Order Fulfillment: The main job of an order fulfillment company is to make sure customers get their products on time and in good shape. A third-party logistics (3PL) company takes care of a business’s whole supply chain. Retailers use these companies for help with keeping stock, storing goods, and sending them out. When you use a 3PL, you don't need your own storage spaces or to manage shipping yourself. You just send your orders to the 3PL, and they handle everything, including picking, packing, and shipping the products.
In the beginning, it's better to outsource and avoid having a warehouse at all. Produce and ship immediately to the customer. This helps to free up money from the goods and use cash flow to produce more or do other things. More importantly it will reveal bad bugs and issues with products - you won’t know they have bugs if they sit in a warehouse.
Make sure you have all necessary and required documents and certification to go through customs clearance.
What else is required post-launch?
Returns: Yes, products will get returned; some will have issues, some customers will just want to return. How do customers return? Which package must they use? Does it require a sticker because it is required for products with batteries by carriers? Where do they send? Who communicates with customers? When do they get a replacement? How do they get a refund? How much does it cost per return? Lots of annoying questions to answer. If you did a good job on design, development, and manufacturing stages, there should be minimal quality related returns.
Repair: Okay, there is a defective product, or some product broke during the warranty. Questions you must prepare to answer: Where do those products get fixed? Who repairs them? How are they trained? Where do they get components? How much do they get paid to repair?
Support: Make sure you have an FAQ in the Manual, and customers know how to find it. Ensure it's available online. Use chatbots to automate support as much as possible. If an answer requires a personal touch, make sure you've appointed someone to deal with support tickets. In the beginning, I did it myself to learn what’s wrong with the product. I strongly recommend you do the same to collect invaluable feedback.
Version 2
While V1 is arriving in the hands of customers, it's time to start work on Version 2 of the product. It's time to evaluate your hypothesis, which features worked, their performance, and the build quality. Identify any weak links in the design. Gather feedback and address issues that users are facing. The goal is to create the shortest feedback loop and capture all feedback. Monitor the usage data. This is a crucial step. Whereas before you were operating based on guesses, now you should have ample data on hand for V2.
Once ready, compare the data with your ideas, decide which changes, features are essential for V2 and move forward.
Next Up
Stay tuned. We'll dive deeper into each step, sharing more insights and tips to help you navigate the path from idea to shelf.
I Want to Hear from You
What's the biggest challenge you've faced in turning your idea into a product? Any specific topics you want to learn more about? Let's get the conversation started.
