"No-code first" - A simple framework for no-bullshit innovation

Why I’m writing this?

You’ve heard about lean startup, agile development, design thinking… Especially if you’ve suffered the pain of waterfall, these new ideas give you hope.?

But, what seems so obvious on paper is so difficult to apply in the real world. You’re paralyzed. You wonder if reading one more book or attending another webinar will unstuck you. Maybe what you need is one of these workshops where you stick tons of post-its in glass walls. I think you don’t.?

By the end of this post, I hope that you see a specific path to start making progress. Progress = seeing a way to test new product and feature ideas at a fraction of the cost you think it actually takes.?

This framework is based on 1 simple idea: frequent action trumps theory. It should be valuable to anyone building, improving, expanding or re-inventing digital products (mobile apps, SAAS, marketplaces, e-commerce…).?

I’m calling this framework “No-code first”. You’ll see why in a sec. Btw, please let me know if someone has used the “no-code first” term before. I’ll happily credit her/him!

Let’s start with a role-playing game.?

We are in the 80s

You are the business leader of the IBM New Ventures team. One of the most promising ideas that the team has been working on is a revolutionary “Speech to Text” product.?

If you have been following the great Alberto Savoia (a big inspiration for this post), you’ll know how this ends.?

This “Speech to text” product is a brilliant piece of tech. Text magically appears on screen as you talk.

Everyone is super excited. Sales forecasts are in the billions. Marketing has been waiting for a product like this for years. Engineers couldn’t be more excited with the technical challenge. Some people even say it will completely replace the keyboard.?

Next week you’ll present the project to get the required funding. You’re going all in.

Someone knocks your office door (we are in the 80s, remember). A few guys from the user research team get in. They are concerned about how end users will perceive the “Speech to Text” solution once they start using it. They suggest testing the product with a few people to find out if they’re right. The test can be done by the end of the week.?

You have 2 options:

1. Ignore them and go with the plan
2. Organize the test and see if they’re right or wrong

Your heart says 1. Your head 2.

Testing day in the lab. 10 people are ready to try the latest version of the “Speech to Text” prototype. As testers start talking, words start appearing on screen. Immediately. No mistakes. First impressions couldn’t be better.?

The team is pumped.?

But the test is not over. As people keep using it for a little longer, some concerns arise. Again and again:

• “We are a few people working in the same space… things would get very noisy if we all used this…”
• “What if I’m working on delicate topics? People around me would hear me. Our legal team won’t like that…”
• “I can see my throat suffering after a few hours…”

The team is shocked.?

During the next few days, you adjust the plan. The company decides to go ahead but you all agree to reduce the size of the bet.?

After launching, the product sees some success but nowhere near the initial expectations. That little experiment might have saved IBM a few hundred millions. This little story is inspired by something that happened for real.?

The key lesson of the story is that the IBM New Ventures team used a simple trick during the testing session in the lab. They faked the whole thing. A fast typist was hiding in another room, typing in real time on a keyboard as testers spoke into the microphone. A highly advanced software product was tested without the need to code anything.?

As you’ll see in a few minutes, for an increasingly big % of cases, new digital product ideas can be tested without writing a single line of custom code. This is why I’m calling this the “no-code first” framework.

Before jumping in, let’s get some context about where this idea comes from.

The feasibility era

A few decades ago, building simple software products was incredibly difficult. Because of that, the number of available software products was small.?

Digital products that were able to solve meaningful problems for the first time became winners. Autocad (1982), Concur (1993) or Salesforce (1999) were huge successes without a fancy user interface or a delightful user experience. During the “feasibility era”, being able to ship mattered more than anything else.?

But, as the number of digital products increased, the % of failed products started to rise too. In 2001, a small group of very smart people, tired of building stuff that made no sense, published the agile manifesto to try to stop the madness. The principles behind the agile manifesto fuel the “no-code first” framework.?

The desirability era

Things started changing in the beginning of the 2000s. First, via development frameworks like Ruby on Rails or Symfony, which dramatically reduced the time needed to go from idea to first functional version of a product.

It continued with the appearance of companies like AWS (hosting), Twilio (communications) or Stripe (payments), which started to offer pieces of functionality that most products need via affordable and easy to integrate APIs. Thanks to these digital building blocks, what used to take months to develop, it now took a few days or weeks.?

The final chapter (until now) has been the recent explosion of no-code and low-code platforms like Bubble, Webflow, Retool or Zapier. They have converted a trend into a revolution.?

For the first time in history, no-code and low-code tools (combined with APIs) make it possible for non-developers to build digital experiences. For an increasing % of products, being able to ship is not a competitive advantage anymore.

This has created a new problem:?

Easier to build → More people building → Many more products in the market

As Naval says:?“It's never been easier to start a company. It's never been harder to build one.”

Data confirms it. According to Harvard Business Review, 30.000 products are launched to the market each year. 95% of them fail.

We’ve become great at shipping but we are still struggling with succeeding. How can we use our new build superpowers to increase our success rate? This is what “no-code first” is about.?

The “no-code first” framework

The “no-code first” thesis can be summarized in 2 sentences:?

1. “The most effective way to discover the potential of a new digital experience is to test it in the real world as quickly as possible.”
2. “The quickest way to deliver a new value proposition via a digital experience is by combining no-code tools and humans.”

Instead of looking at “no-code” as a replacement for code, we treat it as a precursor.?

We use “no-code” to walk the first few steps as quickly and cheaply as possible, always ensuring that we are not serving our users an undercooked solution. Once we know for sure that we’re into something, we consider custom code.

Making it happen

To put the “no-code first” framework in practice, you need 3 things:?

1. An insight
2. A testable idea
3. A desire to find the truth?

1. An insight

An insight is something you know about what a group of people are trying to achieve.

A good insight:

• Needs to be true. You need data (qualitative or quantitative) that proves it.??
• Needs to be commonly understood by everyone involved in finding a solution. Using a written narrative is a good way to do so.?

Good insights come from spending time with your users, customers or prospects. If you have a team of user researchers in place, you should be already receiving a constant stream of insights.?

If you don’t, no worries (nor excuses). You still can get insights. You just need to “get out of the building”. Here’s a great guide about how to do so when you don't have researchers in your team.

A few examples of how an insight looks like:

• “We’ve found that a majority of our customers feel abandoned after they buy a ticket from us. They end up using other apps to fill the needs we are not solving.”
• “We’ve found that some of our customers feel lost when they have to choose between our products. Some of them reach out to our customer service team for help but most just leave our site.”?

2. A testable idea?

A testable idea is a specific offering designed in response to an insight. This simple template will force you to be clear and specific when defining your testable idea.

In the “no-code first” context, “testable” means that you can use no-code/low-code tools to build a first version that you can try-out with real customers.

Most testable ideas can be delivered by combining humans and no-code/low-code solutions. The amount of human involvement needed will vary.?

When testing certain types of AI assistants, humans might have to do most of the heavy lifting. On other occasions, you’ll be able to combine humans with chatbot builders like Landbot.?

For mobile, web or even voice powered apps, you might be able to automate most of the stuff by using platforms like Glide, Bubble or Voiceflow.?

For online marketplaces, you will be able to use platforms like Sharetribe or Softr to present listings and then, if needed, manually connect supply and demand to facilitate the first few transactions.

In most cases, you will be able to use tools Zapier or Integromat to automate repetitive tasks and coordinate everyone involved.?

3. A desire to find the truth

Now that you have a few people using your first version of the solution, you need to find out how much they like it. Or how much they don’t.

The more you’ve been involved in its inception, the more tricky it will be to dodge a dangerous bias: your desire for your testable idea to succeed.?

This is the hardest step in the whole process. You need to find a way to objectively assess what’s happening, not what you’d like to happen. And you need to do so with limited data, since you will likely count your users by dozens.?

I like to focus on two complementary parameters:

a) A user satisfaction metric. This reflects how users rate your testable idea after they’ve used it. Whether you use NPS or the product/market fit survey, you need a metric you can track over time and, even more importantly, something that helps you regularly collect user comments (qualitative feedback).??

b) A retention metric. It measures for how long users keep using the product. This metric needs to be established based on how frequently it makes sense for your users to use your testable idea. A mobile game could use “daily active users (DAU)”. A marketplace like Airbnb could require “repeated bookings in a 6 month period”.????

By tracking these two indicators, you can balance what your users say (satisfaction metric) with what they do (retention metric).?The higher they are, the closer you will be to achieve initial product/market fit.

Disclaimer

The “no-code first” framework doesn’t work for all kinds of digital experiences.

No-code/low-code platforms still lack the level of customization that custom code offers. Sometimes, you’ll need to build a custom solution. This is especially true for API first products like Segment or Bannerbear.

Still, the no-code/low-code industry is rapidly evolving. The % of times where no-code/low-code won’t be enough should be smaller over time.

It’s already happening

“No-code first” is not new. It’s already happening.?

• Bootstrapped startups like Flexiple are building a $3 millions start-up with a no-code stack that costs $60/month.?
• Startups like Edify used only no-code tools to validate their concept ($500k in 3 months) before going all in.
• Successful serial entrepreneurs like Gagan Byani (Udemy, Maven) are systematically using a “no-code first” approach to test their business ideas.
• Ecommerce brands like The Tie Bar are successfully using tools like Typeform to test guided shopping experiences before investing in fully customized solutions.?
• The On Deck No-code chapter is fully dedicated to founders using no-code/low-code to launch new digital businesses in just 8 weeks.
• Here’s a twitter thread with 30 more examples of companies and the no-code stacks they use.?

The “no-code first” mindset is not new but I think it’s time for it to cross the chasm. Hopefully, this post will help a few more people jump on board. Our terrible success rate at launching new products depends on it!

Over to you

What do you think? Can you apply the “no-code first” framework to your current project? Do you know someone who might be interested in it? Feel free to share this post with her/him if that’s the case ;)

Cesc