Essays and guides on startup marketing for seed and Series A founders, most of them written from conversations on my podcast, Founders' Marketing Compass. Newest first, and also by topic. The terms they keep using, from CAC to GEO, are defined in the glossary.
Featured guide
The comprehensive guide to GEO: how to get AI engines like ChatGPT, Gemini and Perplexity to cite you, recommend you and weave you into the answer. From the fundamentals to real implementation.
At seed and Series A, what is missing is usually a decision: what you sell and to whom. Four ways to fill the gap, how each is paid, what you keep afterwards, and when each one fails.
A quiet change in the Facebook and Instagram in-app browser: tap an ad, and the browser jumps straight to the form and offers to fill it from your profile. Leads come in faster, and in many cases they're worse. And advertisers have no off switch.
Another aggressive Meta Ads default: duplicate an ad with no catalog and the copy comes back with one attached. Didn't check? You may be promoting products you never chose.
How to connect Google Ads, Google Analytics, Meta and TikTok to Claude Code over MCP, read-only first: 14 parts, from setup to cross-platform analysis, with over 100 prompts.
Connecting AI to the real data in your search campaigns, through MCP, turns campaign planning from guesswork into data-driven work. I tested it on three different engines and got three very different approaches.
The stunt is the easy half. Dramatize your sharpest difference cheaply, the way one startup mailed bitten iPhones at the 30% app-store cut, then broadcast it.
Daniel Ziv of Glilot Capital Partners sees hundreds of startups a year. He says founders almost always budget marketing too low, and are too modest to broadcast the wins they do earn.
Until now, Meta's Website Highlights showed a preview of your landing page inside the ad. Now it also adds links to other pages on your site, called Discovery Links. Some of your visitors will land on those pages instead of the campaign's target page. It is on by default, hidden collapsed on the first screen, and you have to check that every link is relevant.
Meta told advertisers it is removing the ability to choose where your ads appear and where they don't. Until now you could say "don't show me here." Soon the automated system decides on its own. There is no date yet, but the notices are already going out. And this is not one small change, it is another example of a bigger trend: our control as advertisers keeps shrinking, and the AI takes it over.
The report from Meta and Google says the ads worked. The one honest test is switching every paid campaign off for a while and reading the growth that stays.
Marketing leaders quit when a startup is built to run on one hero instead of a system. Why the cult-of-personality win breaks, and how to keep the next hire.
Screen for attitude first, then effort, then competence. The interview tell of a bad attitude, why a good hire's skills catch up, and where the rule stops.
Prove a product at home and the target market still won't buy on trust. Put the CEO where the market is; with a real gatekeeper, presence can save years.
Replace every marketing hire you like, and a product the buyer does not want still will not sell. How to tell the two apart, and the test to run first.
A very high marketing return is a sign of room to spend more, not a job done. How to read the number, keep raising spend to its floor, and test brand.
For a consumer app, install costs run into dozens of dollars, and messaging channels cut that toward zero. A low number is only good if users pay it back.
Your first marketing hire works or fails on two answers you settle before day one: the single number marketing will own, and the money you put behind it.
Enterprise buyers in a new market already know your field. A few small, hours-long executive dinners with the real decision-makers beat any launch campaign.
Above 20% of budget at seed or 30% at Series A, marketing spend reads as a red flag to investors. The stage benchmarks, and why a low number can be strong.
Avoid vanity metrics by counting only what serves this quarter's goal. The visitor-count trap, the question to ask of every initiative, and why goals move.
Pay a marketing team on one number it shares with sales, not on leads it counts alone. Why a double quota beats pay-for-performance, and where it breaks.
A founder whose first company collapsed when she left on why you have to be the face early, the point where that turns into a risk, and how to build past it.
Say what you save the customer in one sentence, before the features. The bottom-line test an early-stage investor uses, and the buyers who quietly walk away.
The clearest signal this early is an unpaid reaction that amplifies itself, the way a Reddit thread once did for Honey, well before your metrics can say.
A lead can match every filter and still skip the call. Count the meeting the person attends, get sales and marketing to agree on it, and the no-shows surface.
A startup CMO's short tenure is usually the CEO's to prevent. Why the eighteen-month clock is mostly a founder who never said out loud that it was failing.
A good net dollar retention rate depends on who you sell to: 110-130% for small business, well above 150% for enterprise, and cut any segment below 100%.
Set each marketing target from the number the company is judged on that quarter, then break it to the week. How the cascade holds, and where it breaks.
The strongest marketing lead a startup ever has can be an engineer with no marketing background. When experience matters, and what to screen for instead.
Investors grade a marketing team on three things before any campaign result: one story across product and marketing, a fitting hire, and real tracking.
The clearest red flag in hiring a startup marketing leader is one who wants months to plan before trying anything. What to ask in the interview instead.
A boring category run by lazy, high-margin incumbents is an opening, if you say one plain benefit loudly. What Dollar Shave Club got right, and the test.
The cost Facebook and Google report per customer leans on your cheapest early wins. How to tell harvest from new demand, and model the real number instead.
A return policy can outsell a campaign when the product backs it. How a six-month wear-and-return offer became a store's whole marketing, and how to copy it.
Track the one KPI that tells your business's story at its stage, and read a longer list as a sign you have not found it yet. Shaul Olmert on why one wins.
A 3:1 LTV to CAC ratio usually rests on an undercounted CAC. Count the labor, the software and your own hours, and a real business often sits nearer 1:1.
No, but someone has to. Put a landing page behind a hundred dollars of ads before you hire, read who tries to buy, and know where the cheap test stops.
A high churn rate and fast growth can share one cause: the customers you chase hardest. How an investor traced a company's churn, and the split to run first.
You usually know before you admit it. The one-sentence test for whether a marketing hire is done, the dread that keeps you waiting, and what a quarter costs.
I help seed and Series A founders run their marketing, week by week.