The Fastest Way to Validate Business Ideas
All that’s standing between you and building a billion-dollar company is knowledge. So, why not learn as fast as you can?
👋🏽 Welcome to A Founder’s Life for Me! I’m Alek, and I’ll share my experiences building tech companies to provide practical recommendations on building your own thing.
Subscribe to get free access to all past and future posts.
Why I’m writing about this now
In my last newsletter, I discussed how you can quickly address value risks (the risk that people won’t buy your product) through marketing and sales experiments. The minimum requirements to do this were:
Materials explaining your product and pricing (website or slideshow)
A way for customers to express interest in buying your product (waitlist, email address, phone number)
A way to get your target customer to those materials (outreach or marketing)
If you’re already running a test like this, that’s fantastic. Only about 1 in 10 ideas succeed, so you’re saving yourself tons of time by validating the idea before investing your valuable time and energy. Most people don’t do this. They build a solution in search of a problem instead of finding a problem in need of a solution.
So, where do we go from here? You are already being 10x more efficient by validating ideas before you execute them. Well, today, I’ll explain how to validate 10x more ideas. That’s right, we’re moving at 100x speeds now!
Blinding you with science.
When a scientist runs an experiment, typically, they run more than one at a time. Imagine you are trying to make an optimal cup of coffee in your home French press. Let’s say you start by playing with brew duration and the number of coffee scoops.
If you want to test between 4 and 10 scoops (7 different scoop amounts), and brew times between 330 and 600 seconds (10 different timings, in 30-second increments), you’d be left with 70 experiments to run. If you run one experiment per day, it’d take you 70 days to try every combination.
A scientist would accelerate their testing by, for example, running all of the brew duration tests for a given number of scoops on the same day. Without a lot more work, the scientist is now running experiments 10x faster and has found their favorite cup of coffee in 7 days instead of 70.
Other than my love for coffee and science, why are we talking about this? Well, if you are validating business ideas one at a time, then you’ll be beaten by the person who found an easy way to validate 10x more ideas than you. They’re finding ideas that work in 7 days instead of 70. See, that came together in the end.
Running product experiments like a scientist.
I’m quickly running product experiments through LinkedIn Advertising campaigns. SolidlyAI is an AI meeting assistant designed for working professionals, so I decided to use LinkedIn as my starting platform. You can use whatever platform you think is most relevant to your business idea.
Experiment #1: Who are my customers?
I started by running a LinkedIn Ad campaign to see what type of customers respond to my product. Honestly, the ad wasn’t very good (see below), but it served its purpose to see who would respond.
LinkedIn provides insights into who is responding (clicking on) your ad, and the type of company they work for. This is measured by “click-through rates” (CTRs), which tell you, “For every 100 people that see your ad, what percentage of them will click on it?” The top three job types were:
Sales & Business Development
Customer Success and Support
Project Management
From this, I gather that the people who care about SolidlyAI are primarily in client-facing roles. LinkedIn also provided the industries that these people were in:
Technology and Information Services
IT Services and Consulting
Media and Telecommunications
On top of that, most people who responded to the ad were in entry-level and senior roles. From these three insights, I crafted a target customer profile who was interested in SolidlyAI, “Junior-to-mid level employees in client-facing roles at technology and consulting companies.”
Now, this may sound like one experiment, but this single advertisement was served across 25 different industries and 25 different job functions. So, behind the 3 that worked, there were 22 failed experiments for targeting job functions and 22 failed experiments for targeting industries. 50 total experiments, a lot of insights, and no building required.
Experiment #2: Tailoring SolidlyAI to my target customer.
With the insights from my first experiment, I created a new webpage focused on the client management audience. And, six new advertisements focused on messaging to that audience. I launched a new campaign focused only on that audience. Sure enough, I ran this campaign and achieved higher CTRs.1
The advertisement above was a more general “baseline ad,” and, in parallel, I tested five other problem statements that, I hypothesized, might matter to this audience. These hypotheses came out of my own experience, combined with conversations I had with people that fit my target customer profile from experiment #1:
“Project transitions from sales to customer success and support teams are hard.”
“Keeping the whole team up-to-date on the latest client communication is hard.”
“Staying organized is difficult when I have multiple clients.”
“Taking care of follow-ups is hard when I’m in back-to-back meetings all day.”
The 5th was a bit of a wildcard. I took a “leverage AI to securely augment your client support” angle to see if the “security-first” angle would work. (It didn’t.)
I found that the most effective ads (by CTR) were:
“Staying organized is difficult…” (2.6% CTR)
“Taking care of follow-ups…” (2.4% CTR)
“Keeping the whole team up-to-date…” (2.4% CTR)
“Project transitions from…” (2.2% CTR)
Baseline ad shared above… (2.2% CTR)
“Securely leverage AI…” (1.4% CTR)
These CTRs were, on average, 20% higher than the more general ad from my first experiment, which averaged a 2% CTR. Also, for reference, the average LinkedIn Ad has a 0.5% CTR.2 So, I’m now reaching a 5x above-average problem to solve.
So, with this experiment, I learned:
The target audience (identified in experiment #1) was the right audience
Proven through the measurement of higher CTRs and lower advertising costs required for this audience vs. the “general audience” targeted in experiment #1.
What matters most to this audience is 1) staying organized across multiple clients, 2) taking care of follow-ups, and 3) sharing updates with teammates
Proven through the comparison of CTRs and advertising costs across the six problem statements tested in experiment #2.
With this method, I got to stack rank customer problems and focus on what matters to my target customer without building anything new. If I hadn’t done this and just started building a product designed solely to “manage project transitions,” I’d immediately be at a disadvantage relative to focusing on a product that “helps people stay organized when working with multiple clients.”
For people who have an idea and aren’t doing this.
“Running ads is too expensive” - I was able to get to the insights above with $200 spent for each experiment. If your time is worth $20/hour, then you only need to save yourself 10 hours to break even on the $200 investment. That means you’re breaking even if this saves you from going in the wrong direction for one day. If you don’t have $200 to spare, then you can run smaller-scale experiments for $25 or $50; you just can’t validate as many ideas in parallel.
“I don’t know anything about running digital ads” - I had literally never created an ad before the first one you saw above, and it was all fairly straightforward. If you can’t figure it out, send me an email at newsletter@alekhagopian.com, and I can help you.
“This will take too much time” - Setting up my first experiment took me less than a day. Setting up my second experiment took me less than half a day (and it was more complicated than the first). Setting up future experiments like these has continued to take me less than half a day. You can add this to the “running ads is too expensive” argument, and it’ll usually still be worth it.
“I don’t know how to interpret the results” - I built a small Google Sheets calculator to help determine whether the differences between the campaigns were meaningful enough for me to draw conclusions. If you’re struggling to understand the results of your campaign, send me an email at newsletter@alekhagopian.com, and I can help you.
“But, I don’t want to bug you with my stupid questions” - A big reason that I’m doing these newsletters is so that you can be less intimidated by the unknown. So, feel free to reach out. This help is free. I don’t intend to sell you anything. Collectively, thousands of talented people spend millions of hours every week on things that other people don’t want. I’d love to put a small dent in those numbers by offering my time.
Building something people want.
Again, value risk is “whether customers will buy it or users will choose to use it.” Running ads allows you to quickly understand and reduce your value risk. The more you can parallelize experiments, the faster you can reduce your value risk and the more likely your company is to succeed.
Recommendation: For your next business, product, or feature idea. Run an ad campaign for it before you build it to validate that it solves a real problem. Parallelize many campaigns to learn as much as you can as quickly as you can.
All that’s standing between you and building a billion-dollar company is knowledge. The biggest knowledge gap for most founders is “not knowing what customers want.” So, the quicker you learn what customers want, the quicker you and your company will achieve greatness. Running rapid experiments through advertising campaigns can help you learn 100x faster than a traditional “build then market then sell” approach. So, why not step on the gas?
If you are interested in more content like this, subscribe through the link below.
In parallel to the client-oriented language, I also ran a slightly updated version of the old advertisement to make sure that the difference I was measuring wasn’t just due to the ads running at different times of the year.
If you have any ideas for topics you’d like to see me cover, let me know! Next week, I’ll be taking a suggestion from a reader and talking about setting goals for 2024.