Practical guides for founders: find the right problem, build the minimum, win customers and understand your numbers. No fluff.
Validation & customers
How to find market problems that are worth solving
Tank AI Team9 min readValidation
Most businesses don't fail because of bad execution. They fail because they solved a problem nobody felt strongly enough. Before you build anything, your number-one job is to find a problem that is real, frequent and costly for someone who can pay.
In 30 seconds
A good problem is frequent, painful and expensive to ignore.
Look for painkillers (they relieve urgent pain), not vitamins (nice to have).
Validate with behavior, not opinions: what people do > what they say.
The right problem has people already spending time or money solving it badly.
Vitamins vs. painkillers
The first mental filter is simple: is your idea a vitamin or a painkiller? Vitamins are pleasant ("it would be nice to have this"), but nobody changes their routine or opens their wallet for them. Painkillers solve a present, concrete, annoying pain; people actively look for them and pay to stop hurting.
The trap is that almost every idea sounds interesting inside the founder's head. The useful question isn't "would anyone like this?" but "is there someone so uncomfortable with this that they're already trying to fix it somehow, even with a spreadsheet, by hiring someone or by paying for a patch?"
The three dimensions of a good problem
A problem worth solving almost always meets three conditions at once:
Frequency: it happens often. A once-a-year pain rarely justifies a product; a weekly or daily one does.
Intensity: when it happens, it really hurts: it costs money, time, reputation or peace of mind.
Willingness to pay: someone with budget and authority wants it solved.
If a problem is frequent but painless, it gets ignored. If it's intense but extremely rare, there's no recurring market. Businesses live where all three intersect.
The best problems live in the top right: they happen often and hurt enough that someone will pay to solve them.
Where to look for real problems
You won't find good problems brainstorming at your desk. They show up where people already complain or improvise:
Your own experience: tasks you repeat and hate, broken processes in your job or industry.
Workarounds: when you see someone solving something with spreadsheets, WhatsApp or by hiring another person, there's a pain without a product.
Communities and reviews: forums, groups and competitors' 1-2 star reviews are goldmines of real frustration.
Legacy industries: sectors where "it's always been done this way" tend to hide huge inefficiencies.
Validate with behavior, not opinions
"Would you use this?" is the most dangerous question in entrepreneurship, because people are polite and will say yes. What matters is costly behavior: is the person already spending time, money or energy trying to solve the problem? Do they join a waitlist, book a call, share private information or pre-order?
A good problem interview doesn't ask about the future ("would you buy?") but about the past ("tell me about the last time this happened: what did you do, how much time did you lose, what did it cost you?"). The past is evidence; the future is opinion.
A quick test: the money question
If you describe the problem and the person sighs and, without being asked, tells you what they've already tried to fix it, you're on the right track. If you have to explain why they should care, they probably don't care enough.
Where Tank AI fits
Tank starts exactly here: before building anything, its research agent helps you map who has the pain, how often it happens and how much it costs today, so you don't build a vitamin disguised as a painkiller. It's the difference between "I have an idea" and "I have a validated problem."
What an MVP is (and the most common mistakes people make about it)
Tank AI Team8 min readProduct
An MVP isn't a cheap, ugly version of your dream product. It's the smallest experiment that tells you whether it's worth continuing. Confusing the two is one of the most expensive mistakes in entrepreneurship.
In 30 seconds
MVP = Minimum Viable Product: the minimum needed to deliver value and learn.
"Viable" means it actually solves something, not that it's unfinished.
The goal isn't to launch fast: it's to learn fast with the least effort.
Classic mistake: building half a house instead of something small that already works.
What an MVP really is
The working definition: an MVP is the smallest version of your product that already delivers value to a real user and lets you validate a key business hypothesis. The word people ignore is "viable": it has to work and solve something, even if it does very little.
An MVP answers questions like: do people want this? Do they use it more than once? Do they pay? Do they recommend it? If your first version can't answer any of those questions, it isn't an MVP: it's a demo.
The half-a-house mistake
The most useful metaphor for the MVP is transportation. The wrong way to build a car is to deliver a wheel first, then a chassis, then a body: for months the user has nothing to use, and only at the end do you find out whether they wanted a car. The right way is to deliver something that already gets them around (a skateboard, then a bike, then a motorcycle), improving with each step. At every stage the user gets somewhere and you learn.
The right MVP is usable at every step; the wrong one only works at the end, when you've already spent everything.
The most common mistakes
1. Confusing "minimum" with "mediocre"
Minimum refers to scope (few features), not quality. An MVP can do just one thing, but it has to do that thing well. A bad experience gives you false data: people won't be rejecting your idea, they'll be rejecting your execution.
2. Building before validating the problem
The MVP validates the solution, not the problem. If you didn't first confirm the pain exists (see the article on finding problems), your MVP will only tell you that you built something nobody needed, and built it well.
3. Adding "just one more feature"
Feature creep kills MVPs. Every extra feature delays learning and dilutes the signal. The right question before adding something is: "do I need this to test my main hypothesis?" If the answer is no, it goes to the backlog.
4. Not defining what you're going to learn
An MVP without a hypothesis is just work. Before building, write down: "I believe [type of user] [will take this action] because [reason]; I'll know if [metric]." Without that, you won't know whether the experiment succeeded or failed.
The MVP that isn't a product
Often the best MVP has no code. A landing page with a "buy" button that measures clicks, a pre-sale, a service done by hand behind the scenes (the "Wizard of Oz"), or a PDF sent by email can validate the hypothesis faster and cheaper than an app. Remember: the goal is to learn, not to impress.
Where Tank AI fits
Once the problem is clear, Tank helps you set up that first experiment (storefront, offer and payments) in its minimum viable form, so you can test whether people pay before investing months in building. Less "half a house," more real learning.
First-principles thinking and how to apply it to your business
Tank AI Team9 min readMindset
Thinking from first principles means breaking a problem down to its most basic truths and rebuilding from there, instead of copying what everyone else does. It's the difference between making the candle 10% better and inventing the light bulb.
In 30 seconds
Reasoning by analogy = copying what exists. By first principles = building from the fundamentals.
The method: question, break down, rebuild.
It works for pricing, product, costs and business model.
It's not thinking more; it's thinking from further down.
Analogy vs. first principles
Most decisions are made by analogy: "others charge X, so I'll charge X"; "every SaaS has these features, so I'll have them too." It's efficient, but it locks you inside the limits of what already exists. First-principles thinking does the opposite: it ignores convention and asks "what do I know to be true here, and what follows from that?"
The famous example is batteries. "They've always been expensive" is reasoning by analogy. Breaking them down into their materials (lithium, nickel, cobalt, casing) and calculating the real cost of each revealed they could be much cheaper than the market assumed. The truth was one level below the convention.
The three-step method
From convention to basic truth, and from truth to a solution nobody had considered.
How to apply it to your business
In the problem
Instead of "I need an app like X," go one level down: "what result does my customer really want?" You may find they don't need your app, but the result you promise, and that there's a more direct way to deliver it.
In pricing
Don't copy your competitor's price. Break down the value you create (time saved, money earned, risk avoided) and the real cost of delivering it. The price comes from that truth, not from the competition's price sheet.
In costs
"Running this is expensive" is often an assumption. Break each cost into its parts. Often 80% of the spend sits in one step you could automate, eliminate or renegotiate.
Watch out for false truths
The method only works if your "basic truths" really are truths and not opinions in disguise. "My customer doesn't care about price" isn't a first principle; it's a hypothesis you need to validate. The discipline lies in separating what you know (with evidence) from what you assume.
Where Tank AI fits
Tank is built on this way of thinking: instead of giving you "one more landing page," it breaks your business down into its real pieces (customer, pain, offer, payments, delivery) and rebuilds a complete economic unit from there. First principles applied to building companies.
How to get organic visibility for your business (without burning money on ads)
Tank AI Team8 min readGrowth
Organic visibility is the attention you earn without paying for every click: content, community, SEO and word of mouth. It's slower than ads, but it compounds over time and doesn't switch off when you stop paying.
In 30 seconds
Organic compounds: every piece keeps working for you for months.
Pick one or two channels where your customer already is, not all of them.
Create content that solves, not content that only promotes.
The flywheel: content → audience → trust → demand → data → better content.
Why organic wins over time
Ads are a faucet: turn it on, traffic flows; turn it off, it stops. Organic is an asset: a good article, a useful video or a valuable answer in a community keeps attracting people months or years after you publish it. At first organic feels slower (and it is), but compounding works in your favor.
The visibility flywheel
Organic visibility isn't a list of disconnected tactics; it's a cycle that reinforces itself. Think of it as a flywheel: it's hard to get moving at first, but every turn makes it spin more easily.
Each stage feeds the next: content attracts an audience, the audience builds trust, trust creates demand, and the data teaches you to create better content.
Choose your channels deliberately
The most common mistake is trying to be everywhere. Better: identify where your customer already spends time and master one or two channels before opening a third.
SEO / written content: ideal if your customer searches for solutions on Google. It compounds like nothing else.
LinkedIn / X: powerful for B2B and for building a founder's personal brand.
Communities: answering with real value where your customer already gathers builds trust fast.
Word of mouth: the most powerful of all; you earn it by delivering a memorable product and memorable support.
Content that attracts, not content that interrupts
The golden rule: your content should be useful even if the person never buys from you. Answer the real questions your customer is asking, teach them how to solve their problem, share what you learn. The sale comes as a result of trust, not as the main message. Paradoxically, the less you sell in your content, the more you sell with it.
Where Tank AI fits
Tank's growth team helps you get that flywheel spinning: it defines the channels where your customer is, proposes content and sequences, and measures what attracts the best demand, so organic visibility stops being a "someday" and becomes part of how you operate.
Finance basics for first-time founders: P&L, cash flow and balance sheet
Tank AI Team10 min readFinance
You don't need to be an accountant to run a business, but you do need to understand three reports: the income statement (P&L), cash flow and the balance sheet. They tell three different stories about your money, and mixing them up can cost you the company.
In 30 seconds
P&L: did I make or lose money in a period? (profitability)
Cash flow: how much cash actually came in and went out? (liquidity)
Balance sheet: what do I own and what do I owe today? (solvency)
You can be profitable on paper and still go under for lack of cash.
Three reports, three questions
The most expensive confusion for a first-time founder is believing that "having sales" means "having money." It doesn't. Each report answers a different question, and you need all three to see the full picture.
Same business, three views: it can be profitable (P&L) and still run out of cash (cash flow) in the same month.
1. The income statement (P&L)
The P&L (profit and loss, or income statement) measures your profitability over a period: a month, a quarter, a year. The basic formula is: revenue − costs − expenses = profit. It tells you whether your business model makes money on paper.
Key concepts: gross margin (revenue minus the direct cost of what you sell) tells you how much you have left to operate; net income is what's left at the end, after everything.
2. Cash flow
Cash flow measures the real cash coming in and going out, regardless of when it was "earned" on the books. This is where the number-one killer of small businesses lives: you sell, you're profitable, but your customer pays you in 60 days while your suppliers collect in 15. On the P&L you're winning; in your bank account, you have nothing left to operate with.
That's why, for a young company, cash flow is more urgent than profitability. Profitability is an opinion about the future; cash is today's reality.
3. The balance sheet
The balance sheet is a snapshot of your position at a single moment: what you own (assets), what you owe (liabilities) and what's left for you (equity). It always holds: Assets = Liabilities + Equity. While the P&L and cash flow cover a period, the balance sheet is a snapshot of your solvency on a given date.
How to start without overcomplicating it
You don't need accounting software on day one. Start with a simple spreadsheet: record every cash inflow and outflow (your cash flow), add up the month's revenue and costs (your P&L) and, once a quarter, list what you own and what you owe (your balance sheet). What matters isn't sophistication but consistency: up-to-date numbers that you actually look at.
Where Tank AI fits
Tank tracks your business economics from the first dollar: revenue, margins, MRR and cash flow, integrated with your payment gateway. So you don't discover a cash problem when it's too late; you see it on your traction dashboard while you can still act.
Most customer interviews produce polite lies. Brian Long’s "problem hunting" method (he founded Attentive) flips that around: instead of asking about the future, you hunt for real problems by digging into the person’s recent past.
In 30 seconds
Don’t ask "would you use it?": ask about the last time the problem happened.
Hunt for problems, not validation of your idea.
Keep asking "why?" until you reach the root cause and the emotion.
A strong signal: the person has already spent time or money trying to solve it.
Why normal interviews fail
When you ask "would an app that does X help you?", people want to be nice and almost always say yes. You get a compliment, not data. Problem hunting starts from an uncomfortable premise: people don’t know what they want, but they do remember what frustrated them. Your job isn’t to pitch your solution; it’s to uncover pain that already exists and measure how much it actually hurts.
The questions that actually work
The core of the method is anchoring the interviewee in concrete facts from the past, not hypotheticals:
"Tell me about the last time X happened to you." The past is evidence; the future is opinion.
"What did you do to solve it?" If they did nothing, the pain is mild.
"How much time/money did it cost you?" Quantify the intensity.
"Why was that a problem?" Repeat it until you hit the root cause.
Move the conversation from the hypothetical future to the concrete past: that’s where the truth lives.
How to read the signals
A value interview ends with an honest classification. Did the person describe frequent, intense pain, or were they just being polite? Have they already tried to solve it (strong signal), or did they only notice the problem when you brought it up (weak signal)? The goal isn’t to hear "I love it"; it’s to find someone so frustrated they lean forward when you describe the problem.
Common mistakes
Talking more than the customer, pitching your idea too early, asking questions that hint at the answer, and settling for generalities ("it’s annoying sometimes") without asking for the specific case. The discipline is staying quiet, asking "why?" one more time, and resisting the urge to sell.
Where Tank AI fits
Tank’s research agent helps you structure these conversations: who to interview, what to ask and how to classify the signals — so you only start building once the pain is proven, not assumed.
"Entrepreneurs" or "businesses" isn’t a customer: it’s a map of the whole world. Your ICP (Ideal Customer Profile) is the sharp description of who you serve best — and choosing it well is the decision that organizes everything else.
In 30 seconds
A vague ICP dilutes your message, your product and your sales.
Define who decides, who pays and who uses it — they’re not always the same person.
Include the buying trigger: what has to happen for them to go looking for you.
Better to be essential to a few than irrelevant to many.
Why a sharp ICP wins
When you try to serve everyone, your copy turns generic, your product fills up with features for no one and your sales team doesn’t know who to chase. A specific ICP does the opposite: it lets you speak one person’s exact language, prioritize the features that matter and focus your acquisition. Paradoxically, the narrower your ICP, the faster you grow, because you dominate one segment before expanding.
The dimensions of a good ICP
Demographics aren’t enough. A useful ICP combines several layers:
Pain and urgency: what problem they have and how urgent it is.
Decision roles: decision-maker, payer and user.
Ability and willingness to pay.
Trigger: the event that sets off the search for a solution.
Reachable channel: where you can find them.
From the total market to the core: your ICP is where pain, ability to pay and a real trigger intersect.
From a generic ICP to a specific one
Compare: "SMBs that need marketing" vs. "design agencies of 5–15 people in LatAm that lose projects because they have no sales process, and just hired their first salesperson." The second version tells you what to build, what to say and where to look. Start narrow; you can always widen later.
Where Tank AI fits
Tank forces the right question from the start: who exactly is this for. Its agent turns your idea into an actionable ICP with pain, trigger and channel — the foundation it uses to build your offer and your pipeline.
How to validate an idea without writing a line of code
Tank AI Team8 min readValidation
Building is the most expensive and slowest way to learn. Before you code anything, you can test whether people want your idea —and pay for it— with pages, conversations and a bit of manual work. Learning cheaply is a superpower.
In 30 seconds
The goal isn’t to launch: it’s to get behavioral evidence.
A landing page + payment button measures real intent, not opinions.
The "Wizard of Oz" trick: you deliver by hand what looks automated.
A pre-sale is the most honest validation: money up front.
The no-code validation ladder
Not all tests are worth the same. Climb them in order of commitment: each rung costs the user more and, because of that, gives you a more reliable signal.
Each rung asks more of the user and, because of that, better predicts whether your business is real.
The tactics, one by one
Landing page with intent
A page that describes the offer and a "Buy" or "Get started" button. Measure how many people click. If nobody clicks, the message or the problem is failing — and you found out in days, not months.
Wizard of Oz
The user thinks they’re interacting with an automated product, but behind it you’re doing the work by hand. You validate demand and the flow without building the engine.
Pre-sale
The ultimate test: asking for money up front for something that doesn’t exist yet. If people pay, you have a business; if not, you saved yourself from building it.
Where Tank AI fits
Tank sets up exactly this minimal experiment —storefront, offer and checkout— so you measure real intent before investing in product. It’s validating with behavior, not surveys.
How to talk to users without biasing their answers
Tank AI Team7 min readValidation
Your users want you to like them, and that contaminates everything they tell you. Learning to interview without leading the answer —the spirit of "The Mom Test"— is the difference between real data and useless praise.
In 30 seconds
Talk about their life, not your idea.
Ask about past facts, not future opinions.
Avoid questions that hint at the answer you’re hoping for.
Compliments are noise; commitments are signal.
The golden rule
The name "The Mom Test" comes from a simple idea: ask questions so good that even your mom —who loves you and wants to cheer you on— can’t lie to you. How? Don’t tell her your idea. If she doesn’t know what you’re selling, she can’t tell you what she thinks you want to hear. You talk about her life and her problems, not your solution.
Good questions vs. bad ones
The difference is whether the question can be answered with a compliment or requires a fact:
❌ "Would you like an app for this?" → invites praise.
✓ "How do you handle this today?" → reveals the real process.
❌ "Would you pay $20 a month?" → a hypothetical, polite answer.
✓ "What tools do you pay for today to do this?" → real behavior.
Three filters so the conversation produces truth, not courtesy.
What to do with what you hear
Separate facts (what they do and have paid for) from opinions (what they think they would do). Give weight only to the former. And look for the next real step: an intro to their boss, another meeting, a pre-sale. A small commitment is worth more than a thousand "looks great"s.
Where Tank AI fits
Tank uses this discipline as input: it turns what you learn from your customers into a sharp ICP and offer, instead of nice-sounding assumptions that fall apart at the first sale.
Value proposition: the formula that connects pain to outcome
Tank AI Team8 min readProduct
A good value proposition doesn’t describe your product: it describes your customer’s transformation. If people can’t tell in five seconds what they gain from you, no amount of pretty design will save it.
In 30 seconds
Sell the outcome, not the features.
Connect a specific customer with a pain and an outcome.
Include your differentiating mechanism: why you.
If it sounds like it’s for everyone, it convinces no one.
The formula
A simple, powerful structure for writing it: We help [specific customer] achieve [concrete outcome] by solving [painful problem], without [main friction], through [differentiating mechanism]. It doesn’t have to sound like a slogan; it has to be clear. Clarity sells better than creativity.
Your value proposition is the bridge between "today, with the pain" and "after, with the outcome."
Mistakes that weaken it
Talking about features instead of benefits ("it has AI" instead of "it saves you 5 hours a week"); being so generic it applies to anyone; promising everything to everyone; and using internal jargon your customer doesn’t use. A simple test: show it to someone in your ICP for five seconds and ask what they understood. If they miss it, rewrite it.
How to test it
A value proposition isn’t validated in a meeting: it’s validated on a landing page. Put it at the top and measure whether people keep reading and click. The market will quickly tell you whether you connected pain to outcome.
Where Tank AI fits
Tank writes and tests your value proposition as part of building your storefront: it ties it to your ICP and puts it in front of real users, adjusting it based on what actually gets clicks.
Getting new customers is expensive and addictive; keeping them is what builds a business. Without retention, acquisition is filling a leaky bucket: no matter how much you pour in, it never fills up.
In 30 seconds
Retention is the foundation of a recurring business’s economics.
Without retention, every dollar spent on acquisition is wasted.
Measure retention by cohort, not a misleading average.
The best growth channel is a product people don’t want to leave.
The leaky bucket
Picture your customer base as a bucket. Acquisition is the water coming in; churn (cancellations) is the hole it leaks out of. If the hole is big, you can pour faster and faster and never raise the level. That’s why, before you hit the gas on marketing, you plug the hole: a product that retains turns every acquisition effort into compounding growth.
If you don’t plug the churn hole, all the acquisition water leaks out.
How to measure it right
A global retention average hides the truth. Use cohorts: group customers by the month they joined and see how many are still active 1, 3 and 6 months later. If the curve flattens at a healthy level, you have product-market fit; if it drops to zero, the product isn’t sticking no matter how much you sell.
What drives retention
Getting to the "aha moment" fast (the instant the user sees the value), building habit, and support that actually solves problems. Retention isn’t a feature you add: it’s the result of the product keeping its promise again and again.
Where Tank AI fits
Tank measures retention by cohort from day one and tells you when the hole is the problem — so you don’t burn budget on acquisition into a broken bucket.
Support as product: the service that keeps customers
Tank AI Team7 min readProduct
Support isn’t a cost center to minimize: it’s one of the few moments you talk directly with a customer who has a problem. Done well, it turns friction into loyalty.
In 30 seconds
Support is part of the product, not an add-on.
A problem solved well retains more than if it had never happened.
Every ticket is free feedback on what to fix.
Systematize what repeats; humanize what’s sensitive.
The recovery paradox
There’s a well-documented phenomenon: a customer whose problem gets solved quickly and well often ends up more loyal than one who never had a problem. Support is your chance to prove there’s someone behind the product who answers. Ignoring it, on the other hand, is the fastest route to refunds and one-star reviews.
What to systematize and what to humanize
Not everything should be handled the same way. FAQs, login problems and billing questions can be automated or handled with templates and a good knowledge base. Sensitive cases —a large refund, an upset key account, a bug that cost money— need human judgment and should escalate to the right person with full context.
Automate what repeats; save your time and judgment for what really matters.
Support as a source of product
Every conversation is a clue: if ten people ask the same thing, there’s a missing feature, confusing copy or an unnecessary step. Turn recurring tickets into product improvements and you’ll watch support volume drop over time.
Where Tank AI fits
At Tank, support is part of the unit economics: your team answers the routine questions with the customer’s full context, and the delicate cases escalate to you, the CEO, ready for a decision.
Founders’ #1 mistake: not being clear on the problem
Tank AI Team8 min readMindset
Most founders fall in love with their solution before they understand the problem. The result is a polished product that solves something nobody was losing sleep over. Problem clarity is the compass that organizes everything else.
In 30 seconds
Fall in love with the problem, not the solution.
If you can’t explain the problem in one sentence, you don’t understand it yet.
A clear problem guides product, message, price and channel.
"I’m building X" ≠ "I solve Y for Z"; the second one is a business.
Symptoms of a lack of clarity
You’ll recognize it when: your pitch describes features, not outcomes; your target customer is "everyone"; you change your mind every week about what to build; and when you explain your product, people nod politely but never ask how to get it. They’re all symptoms of the same root cause: you don’t know precisely what pain you solve or for whom.
A well-defined problem is the source every other decision flows from.
How to get clarity
Write the problem in a single sentence: "[customer] can’t [achieve outcome] because [obstacle], and that costs them [consequence]." If you can’t fill it in with evidence, go talk to customers until you can. Clarity isn’t something you think up at your desk: you discover it in conversations.
Where Tank AI fits
Tank starts with the problem, not the product: it makes you define customer, pain and outcome before building, so you don’t end up with a beautiful solution in search of a problem.
How to spot a growing industry and build with the market's momentum
Tank AI Team9 min readStrategy
It's much easier to swim with the current. Starting a business in a growing industry means that even when you make mistakes, the rising tide lifts you. Picking the right market can matter more than flawless execution in a stagnant one.
In 30 seconds
The market is the wind: choose one at your back.
Look for recent shifts that open up opportunities: "why now?"
Typical waves: new technology, regulatory, generational or cost shifts.
A growing market forgives mistakes; a stagnant one punishes them.
The "why now?" question
The best opportunities show up when something changed recently and opened a door that used to be shut. The key question is "why now?": what became possible, cheap, legal or necessary that wasn't a few months ago? If your answer is "you could always do this," someone has probably already done it, or the market doesn't want it.
The triggers (technology, regulation, generational and cost shifts) are the waves that create new markets.
Where to look
Technologies that just became accessible (like AI today), regulatory changes that create new needs, the habits of one generation replacing another, and cost drops that make something once expensive viable. Cross those waves with a real problem and a customer who pays, and you'll have momentum on your side.
Watch out for hype
Growth isn't the same as hype. A market can be trending on Twitter and still have no paying customers. Make sure there's real money changing hands behind the growth, not just enthusiasm.
Where Tank AI fits
Tank's research evaluates your idea's "why now?": market size, maturity and readiness to adopt — so you build where the current pushes you forward, not where you're rowing alone.
Automate vs. validate: when to build and when to wait
Tank AI Team7 min readMindset
AI makes building easier than ever — and that's the danger. When automation is cheap, the temptation is to skip validation and churn out, at lightning speed, things nobody wanted. Knowing when to wait is an advantage.
In 30 seconds
Automate a process only after you understand and validate it.
"Do it by hand" first: learn where the real value is.
Automating the wrong thing just makes you fail faster.
Rule: validate the what, then automate the how.
The risk of speed
Cheap building is wonderful for iterating, but it also lets you scale a mistake at high speed. If you automate a workflow that doesn't solve a real problem, all you'll get is a lot of something useless. The question isn't "can I build this?" but "should this exist?"
By hand first
Before automating, run the process manually a few times. Delivering the service by hand teaches you where the value is, which steps are unnecessary and what the customer really expects. Only once you understand the process — and have confirmed people want it — does it make sense to invest in automating it.
Automation is the last step, not the first: validate the what before you optimize the how.
Smart friction
Sometimes putting a small obstacle in front of building — forcing yourself to land three paying customers first, for example — is the healthiest thing you can do. The right friction keeps you from automating your way to the wrong place.
Where Tank AI fits
Tank builds in that smart friction: it pushes you to validate before you build, and its agents automate only once the process has proven it deserves to exist.
From idea to first dollar: a business's minimum sequence
Tank AI Team9 min readStrategy
Your first real dollar changes everything. Not because of the money — because it proves someone values what you offer enough to pay for it. This is the shortest path to get there.
In 30 seconds
A first payment validates your business more than a thousand "likes."
The minimum sequence: problem → offer → storefront → checkout → first sale.
You don't need the full product; you need a promise people will pay for.
Aim to sell before you build everything.
Why the first dollar matters so much
There's a huge gap between "people like it" and "people pay for it." Money is the only vote that truly costs something. Getting your first payment proves the problem hurts, your offer relieves it and your price is acceptable. Everything before that is theory; the first dollar is the first evidence.
The minimum sequence
Five pieces, in order: each one exists to get to the sale, not to show off.
The shortcut: sell before you build
You don't have to finish the product to get paid. A presale, a paid pilot or a service delivered by hand gets you your first dollar while you're still learning. The ideal order isn't "build, then sell," but "sell, then build just enough to deliver."
After the first dollar
Once someone pays, your job changes: you're no longer validating whether they want it, but whether you can deliver value in a repeatable, profitable way. The first dollar opens the door; retention and margin keep it open.
Where Tank AI fits
Tank is designed around this sequence: it turns your idea into an offer, a storefront and a checkout ready to sell — to cut the path to your first real dollar as short as possible.
SaaS, consulting, e-commerce or a course: which business type to choose
Tank AI Team9 min readStrategy
The same know-how can become a consulting practice, a course, a SaaS or an e-commerce store — but each model has different building blocks, different rhythms and different traps. Choosing well based on what you have today speeds everything up.
In 30 seconds
Not every business is built the same way; each has its own pieces.
Consulting: fast to get paid, depends on you.
Course: scales if you already have an audience.
SaaS: recurring, but demands a product and retention.
E-commerce: the most operational (inventory and logistics).
Four models, four logics
The question isn't "which one is best" but "which one fits what I have and how I want to work." Your expertise, your audience, your tolerance for physical operations and your time horizon make the call for you.
Choose by fit, not by trend: what you already have (expertise, audience, product) tips the scale.
A recommendation to get started
If you have knowledge and not much else, consulting is usually the best first move: it requires no inventory or finished product, you can charge high fees and you learn a ton from the customer. A course works if you already have an audience; SaaS is more ambitious; and e-commerce is the most operationally demanding. Start with whatever you can charge for fastest and evolve later.
Where Tank AI fits
Tank classifies your idea by business type and adapts all execution to its building blocks — because launching a SaaS isn't the same as launching a consulting practice, and your plan should reflect that.
Cold outreach that actually works: from list to first sales
Tank AI Team8 min readGrowth
Cold outreach has a bad reputation because almost everyone does it wrong: generic, mass-blasted messages centered on the seller. Done with judgment, it's still one of the fastest ways to land your first customers.
In 30 seconds
List quality > message quantity.
Personalize with a real trigger, not with "{first_name}."
Talk about their problem, not your product.
The goal of the first message is a reply, not a sale.
Start with the list, not the message
80% of the result comes down to who you write to. A small list of prospects who fit your ICP and have a recent trigger (they just hired, grew, or switched tools) beats thousands of generic contacts. Before you write a single word, sharpen the list.
Anatomy of a good message
A cold message that works usually has four parts: a credible reason you're writing to that specific person (the trigger), a brief mention of the problem you suspect they have, proof that you can help, and a low-commitment call to action (a question, not "book a 30-min demo").
An outreach funnel: every stage filters, and the quality of the first one shapes everything after it.
Follow-up is where you win
Most cold sales happen on the second, third or fourth touch, not the first. A short, respectful follow-up — one that adds something new instead of "did you see my message?" — multiplies your replies. Persistence without being pushy is the key skill.
Where Tank AI fits
Tank's growth agent (ClawGTM) builds your lists, enriches prospects and proposes personalized sequences — so outreach stops being copy-and-paste and becomes real pipeline.
How to build an audience before you have a product
Tank AI Team8 min readGrowth
Launching to an audience that already trusts you is infinitely easier than launching into the void. Building an audience before you have a product gives you distribution, feedback and your first customers on day one.
In 30 seconds
Distribution matters as much as the product.
Share your learning process, not just results.
Building in public turns spectators into future customers.
A small, engaged audience is worth more than a large, cold one.
Distribution, the unfair advantage
There are thousands of good products nobody has heard of. An audience is the difference between a launch that echoes into silence and one that starts with traction. If you start building an audience today — months before you have anything to sell — on launch day you'll already have someone to sell to.
Building in public
The most accessible way to start is to share your journey: what you're learning, what problem you're researching, what decisions you're making. You don't need to be an established expert; you need to be useful and honest. People follow whoever teaches them something and shows them what happens behind the scenes.
An audience is an asset that matures: from strangers to customers, by consistently adding value.
Quality over size
Don't chase followers for vanity. A thousand people who actually have the problem you solve are worth more than a hundred thousand curious onlookers. Speak to your ICP, even if it narrows your reach: a small, aligned audience buys; an indifferent crowd doesn't.
Where Tank AI fits
Tank helps you get distribution going early: content and channels aligned with your ICP, so when your offer is ready you already have people to offer it to.
A landing page isn't a pretty brochure: it's a persuasion machine. Its only job is to take a visitor from "I don't know what this is" to "I want this" — and that depends more on clarity than on design.
In 30 seconds
The first thing people see should say what they get, in seconds.
One goal, one main CTA, repeated.
Talk about benefits and outcomes, not just features.
Reduce friction and add proof people can trust.
The structure that works
Almost every landing page that converts follows a similar skeleton, because it mirrors the order in which a person decides: first they understand, then they get interested, then they trust, then they act.
Top to bottom: understand → get interested → trust → act. Each section pushes into the next.
The mistakes that kill conversion
A clever but confusing headline; too many competing goals (subscribe, buy, follow us, read the blog); talking about yourself instead of the customer; asking for too much information; and including no proof that others trust you. Clarity almost always beats creativity: if people have to think, they leave.
Optimize with data
A landing page is never "done." Measure how many people arrive, how many click the CTA and how many complete it. Test a different headline, a different CTA, fewer fields. Small changes to the main message usually move the needle more than any redesign.
Where Tank AI fits
Tank generates your landing page with this structure and connects it to your metrics: you see how many arrive, click and convert, and it adjusts the message based on what actually works.
SEO for founders: how people find you on Google and in AI tools
Tank AI Team9 min readGrowth
SEO is no longer just about showing up on Google: today it's also about getting AI tools to cite you when someone asks. The good news is that the foundation is the same — useful content that answers your customer's real questions.
In 30 seconds
Answer your customer's real questions better than anyone else.
Clear structure (headings, lists, direct answers) helps both Google and AI tools.
Intent > volume: a few high-intent searches are worth more.
SEO compounds: it's slow, but it never stops working.
How Google and AI tools think
Both want the same thing: to give the best answer to a person's question. Google links to it; an AI summarizes it and, if your content is clear and trustworthy, cites or recommends you. Either way, you win if your page answers a specific question directly, completely and honestly.
A good article works on two fronts: the classic search engine and the new AI answers.
The basics that actually matter
Pick questions with clear intent (someone searching "how to validate an idea" is closer to acting than someone searching "entrepreneurship"). Use descriptive headings, direct answers up top, lists and natural language. Make sure the page loads fast and reads well on mobile. And above all, make it genuinely useful: that's still the best SEO trick there is.
Play the long game
SEO is an investment that compounds. An article can take weeks to rank, but then it attracts visitors for years at no cost per click. Publish consistently, answering your customer's questions, and you'll build an asset that works while you sleep.
Where Tank AI fits
Tank suggests topics and content aligned with what your customer is searching for, optimized for Google and for AI tools — so they find you right when they have the problem you solve.
AI-made product videos that don’t look like PowerPoint
Tank AI Founding Team5 min readFree guide
This is how we made Tank AI’s launch videos: an AI writes the HTML and the animations, and HyperFrames turns them into an MP4. No video editor. Here is the full guide, with the prompt so your AI can do it with you.
Free guide · PDF · 4 pages · in Spanish
The 7 steps, the catalog of 17 transitions, the mistakes that already cost us a render, and a prompt ready to paste into Claude, Cursor or ChatGPT.
A video feels alive when something from scene A survives and becomes scene B. We call that change a boundary.
Paper first: script, main character, storyboard and a boundary map. Code after.
Something new every 0.5–1 s, no still shot longer than 1 s, and text at least 22 px tall in 1080p.
Everything in a single composition, animated with GSAP and rendered with HyperFrames.
How ours turned out
Tank AI’s two launch videos, made with this method (in Spanish). Watch them for the boundaries: the symbol that leaves the logo and lands in the field where you type your idea, the row that becomes a label, the button that becomes a window.
Launch video: what Tank AI is and what it does for you (90 s).Product video: the journey from idea to company, cycle by cycle (90 s).
The idea that changes everything
A video feels like PowerPoint when each scene replaces the previous one. It feels alive when something crosses from one scene to the next: the button you press becomes the window, the row becomes the label, the symbol jumps from the brand into the product. The whole method is about designing those boundaries before animating anything.
The 7 steps, in short
Script and music. About 165 words for 90 seconds, and the beat grid of the song. Everything is timed in beats, not seconds.
A main character. Something that crosses the whole video: your brand symbol, a dot of color or the cursor.
Storyboard. One frame per scene: what the voice says, what the main character does and what happens every half second.
Boundary map. One row per scene change. If a row has no answer under “what survives”, it is still a slide.
One world. A single composition with a camera, not separate clips. If every scene is a clip, nothing can survive the change.
Everything that moves makes a sound. Every transition, click and arrival gets its effect, taken from the same map.
Render and check. A 30 fps draft, check every boundary and measure how many frames are nearly still before the final render.
The detail of every step, the 17 transitions and the prompt are in the guide.
Where Tank AI fits
Tank builds your company with agents, in public. This method is part of how we tell that story: your product, explained in 90 seconds people actually watch.
Pricing for beginners: how to set a price without fear
Tank AI Team8 min readFinance
Setting a price terrifies almost every founder, which is why almost every founder charges less than they should. Price isn't just what it costs to produce: it's a signal of value, and being too cheap can hurt you.
In 30 seconds
Charge for the value you create, not for your costs.
A price that's too low signals "low value" and attracts bad customers.
Lowering prices is easier than raising them: start high.
Price is a hypothesis: test it instead of guessing once and for all.
Cost, competition and value
There are three ways to set a price. Cost-based (what it costs you plus a margin) is safe but leaves money on the table. Competition-based (what others charge) locks you into their logic. Value-based (what the outcome is worth to the customer) is the most profitable: if you save someone $1,000 a month, charging $200 is a bargain for them and a great business for you.
Value-based pricing is almost always the most profitable option — and the hardest one to dare.
The danger of charging too little
A price that's too low doesn't just cut your revenue: it tells people your product isn't worth much, attracts the most demanding and least profitable customers, and leaves you no margin to provide good support. Raising prices later is hard; starting with a respectable price and adjusting is healthier.
Treat it as an experiment
There's no "right" price set in stone. Launch a price, watch your conversion rate and objections, and adjust. If nobody pushes back on the price, it's probably too low. The market will show you where the sweet spot is.
Where Tank AI fits
Tank helps you set and test prices based on the value you deliver, connected to your real payments — so you adjust with conversion data, not fear.
Unit economics answers a brutal question: do you make or lose money on each customer? A business can grow its sales and still be digging its own grave if every customer costs more than they bring in.
In 30 seconds
CAC: what it costs to acquire a customer.
LTV: what a customer brings you over their whole lifetime.
Healthy rule: LTV well above CAC (often 3×).
If you sell more while losing money on every sale, growth sinks you faster.
The three key numbers
CAC (Customer Acquisition Cost) is everything you spend on marketing and sales divided by the customers you acquired. LTV (Customer Lifetime Value) is how much margin an average customer leaves you while they stay with you. Margin is what you keep from each sale after direct costs. Together they tell you whether your growth engine is profitable.
The scale should tip clearly toward LTV: every customer should bring in well more than they cost.
Why the average lies
A business that's profitable "on average" can hide channels that lose money. Measure unit economics by channel and by segment: maybe one channel brings cheap, loyal customers and another brings expensive ones who leave quickly. Shut down what doesn't work and double down on what does.
The growth trap
If you lose money on every customer, growth won't save you: it sinks you faster, because you multiply the loss. Fix unit economics first (raise prices, lower CAC, improve retention) and then accelerate. Growing on healthy numbers is wonderful; on broken numbers, it's fatal.
Where Tank AI fits
Tank tracks CAC, LTV, margin and retention from the very first dollar — so you know whether your engine is profitable before you hit the gas, not after.
Legacy industries: how to build technology products for service businesses
Tank AI Team9 min readModels
"Boring" industries — construction, logistics, repair shops, clinics, agriculture — are full of manual processes and huge opportunities. Where everyone else sees something old, an attentive founder sees a market waiting for technology.
In 30 seconds
Legacy industries have little tech competition and a lot of pain.
The winning model is usually software + service, not software alone.
Start by delivering the service by hand; automate later.
Trust and industry context are worth as much as the product.
Why boring is attractive
Trendy markets are saturated with competitors; legacy industries aren't. Many of these sectors still run on paper, Excel and phone calls. That means big pain, customers with budget and little technology competing for them. The barrier isn't building — it's understanding the industry and earning its trust.
Software with a service glove
In these markets, selling "an app" rarely works: customers don't want to learn software, they want their problem to go away. The model that wins usually combines technology with service: you solve the problem (partly by hand, partly with your product) and the software is the invisible engine that makes it profitable and scalable.
You don't sell software to people who didn't ask for it: you sell the problem solved, with software inside.
How to get in
Start by delivering the service almost entirely by hand for one or two customers; you'll learn the real process and its exceptions. As you understand the patterns, automate what repeats. That closeness to the industry is your edge over a competitor who only knows technology.
Where Tank AI fits
Tank helps you package that software + service model: offer, storefront, payments and operations, with agents that automate the repeatable work while you bring the industry know-how.
Run an agency? How to turn it into a software-as-a-service company
Tank AI Team9 min readModels
An agency trades time for money: every new dollar takes more hours. A SaaS breaks through that ceiling. If you already run an agency, you have something most SaaS founders dream of: real customers and a process that repeats.
In 30 seconds
An agency scales with people; a SaaS scales with code.
Look for the process you repeat for every client: that's where the product is.
Productize first (a standardized service), then automate.
Your hidden asset: you already know which pain your clients pay for.
The agency ceiling
The agency model is great for getting started — you get paid fast and you learn — but it has a ceiling: your revenue is tied to your team's hours. To grow, you hire; margins get squeezed and you spend your days putting out fires. SaaS breaks that equation: you build once and sell many times.
Find the product inside the service
The key is to look at your own operation: what do you do almost the same way for every client? That repeatable process — a report, an analysis, a workflow — is your product candidate. First you productize it (turn it into a standardized service with a fixed scope and price); then you automate the parts software can do better.
The path isn't one big leap: agency → productized service → SaaS, cutting hours at every step.
Your unfair advantage
SaaS founders starting from scratch spend months searching for product-market fit. You already have paying clients and know exactly which pain you solve. That closeness is gold: build the product with (and for) the clients you already have, and you'll land your first subscriptions before you launch to the world.
Where Tank AI fits
Tank helps you make that leap: identify the repeatable process, productize it and build the SaaS — core workflow, login, recurring billing — on top of the clients you already know.
We're writing this resource with the same care as the rest: practical content, real examples and a diagram that explains the concept. Check back soon.
In the meantime
Read the articles already published from the menu.
Got an idea on your mind? You can try Tank AI today.
Where Tank AI fits
Tank helps you turn an idea into a real business: you're in charge, with a team of agents executing alongside you. Explore it while we get this article ready.
