What 375 Campaigns Taught Me About Marketing
375 campaigns. 50+ clients. Some made hundreds of thousands. Some wasted money. Here are the marketing lessons that only come from doing it — not studying it.
Insights
I've run over 375 marketing campaigns across more than 50 clients. Ecommerce brands, service businesses, startups. Founders spending their last $3,000 on ads. Founders scaling through seven figures.
Some campaigns made clients hundreds of thousands. Some wasted money. Most landed somewhere in between.
I've spent a lot of time thinking about what actually separated the ones that worked from the ones that didn't. Not just tactically — which bidding strategy, which creative format — but at the level of thinking and decisions.
The lessons below aren't from courses or frameworks. They're from doing the same thing many times and watching what happened. Some are counterintuitive. Most took multiple expensive failures to actually understand.
1. The Dashboard Is Never the Full Picture
Every platform reports in its own favour.
Meta's ROAS is overstated because it claims credit for sales that Google, email, and organic also claim. Google's conversion tracking fires on sessions that were already going to convert. TikTok counts views as engagements.
None of these platforms are deliberately deceiving you. They're measuring what they can see, through their own lens. But if you make budget decisions based on any single platform's self-reported numbers, you're working with a biased and incomplete picture.
The discipline I use on every client account: pull real numbers first. Total revenue from Shopify or the CRM. Total spend across all channels. Calculate MER. Then look at individual platform data to understand the composition — not the total.
It's slower than reading a ROAS number. It's also accurate.
2. The Things That Feel Like Marketing Often Aren't
Founders spend a lot of time on things that feel like marketing but don't move revenue.
Redesigning the logo. Rewriting the website copy for the fourth time. Debating brand colour palettes. Building elaborate campaign structures before basic messaging has been proven. Creating content for an existing audience of 600 followers that won't reach anyone new.
There's a version of brand work that's genuinely productive. There's also a version that's creative procrastination with a respectable cover story.
The test I find useful: will this activity, in the next 90 days, lead to someone who doesn't know the brand finding out about it — or someone who knows it being more likely to buy? If the honest answer is no, it's probably not marketing. It might be important for other reasons. But it's not marketing.
3. Audience Quality Matters More Than Audience Size
When founders first look at Meta Ads targeting, they're often drawn to the biggest possible audiences. Millions of people. Maximum reach.
The audience that converts best in most campaigns I've run wasn't the biggest. It was the most specific — the tightest match between the product, the problem it solves, and the person experiencing that problem right now.
A cold audience of 50,000 people who match your best customer profile will almost always outperform a broad audience of 5 million, especially early on when the algorithm doesn't yet have enough conversion data to find your buyer inside a huge pool.
The shift toward broad audiences Meta encourages makes sense at scale — when you have thousands of conversions and the algorithm genuinely knows your buyer. Before you have that data, tighter targeting is usually better.
4. Most Campaigns That "Don't Work" Were Never Given the Conditions to Work
I've turned around dozens of campaigns clients told me weren't working. More often than not, the campaigns weren't the problem.
The budget was too low to exit the learning phase. Creative was swapped before any statistical significance. The campaign was paused and restarted multiple times, resetting the algorithm each time. The product page was converting at 0.4% and no amount of ad work was going to fix that.
Good campaigns fail in bad conditions.
When a campaign doesn't perform, the right diagnostic question isn't "what's wrong with this campaign?" It's "have we actually given this campaign what it needs to work?" Those are different questions with different answers. The first leads to endless tinkering. The second leads to fixing the actual constraint.
5. Retention Is Worth More Than Acquisition, but Almost Nobody Acts Like It
The maths of customer retention are well known: acquiring a new customer costs 5-7x more than retaining an existing one. Repeat customers spend more per transaction. They refer more people. LTV, not CPA, determines whether a business model is healthy.
And yet — in the vast majority of new client accounts I've worked on, the email automations are either absent or basic. Post-purchase sequences are missing. Win-back campaigns have never been built. Loyalty mechanics don't exist.
Everyone knows retention matters. Almost nobody has built the systems to act on it.
The reason is that retention work is slower and less immediately satisfying than acquisition work. Running a new Meta campaign has a feedback loop of days. Building a post-purchase email sequence has a feedback loop of months. Founders chase the quick signal.
The clients who've seen the most durable growth are the ones who invested in retention infrastructure alongside acquisition channels. The acquisition brought customers in. The retention system made the unit economics work.
6. The Offer Is Usually the Real Bottleneck
When campaigns underperform, the first instinct is to change the creative or the targeting. These are the visible variables — the things you can directly control in Ads Manager.
But when a brand has a fundamentally weak offer, no amount of creative or targeting work will rescue it.
The offer is: what are you selling, at what price, with what value proposition, against what alternatives? If the price is too high relative to perceived value, if the guarantee isn't there when competitors offer one, if the product page raises more questions than it answers — campaigns will struggle regardless of how good the ads are.
Before I optimise creative, I look at the offer. Before I change targeting, I look at the landing page conversion rate. The bottleneck is almost never where you first go looking for it.
7. Data Shows You What Happened. It Doesn't Explain Why.
Analytics is the most overrated skill in marketing and simultaneously the most underrated.
Overrated because people treat data as a substitute for thinking. A dashboard full of numbers doesn't explain itself. CTR went down — why? Conversion rate improved — because of what? Attributing causality to a metric movement is a skill, not a mechanical readout.
Underrated because most businesses don't collect or structure data well enough to draw any conclusions at all. UTM parameters inconsistent. Conversion events misfiring. GA4 set up but never reviewed.
The discipline is this: clean data infrastructure first. Then build the habit of looking at trends over time. Then — separately — develop the judgment to understand what the data is actually telling you versus what you want it to say.
Numbers show what happened. Figuring out why requires thinking, not more data.
8. Speed of Learning Beats Perfection of Execution
In the early stages of any channel or campaign, the goal isn't to launch a perfect campaign. It's to find out what works as fast as possible.
A good-enough campaign that launches in one week, generates data, and informs a better second version the following week will outperform a perfect campaign that takes six weeks to build. Learning compounds. Delay doesn't.
This is the most important mindset shift for founders who come from non-marketing backgrounds. The engineering instinct — build it right before you ship — will actively slow you down in marketing, where the right answer can only be discovered by running something and watching what happens.
Ship a testable version. Learn. Improve. Repeat. The version you're on in three months will be far better than whatever you'd have built in the first six weeks of planning.
9. The Clients Who Treat Marketing as Infrastructure Win
The most consistent pattern I've seen across 50+ clients: businesses that treat marketing as infrastructure — something to build properly and maintain, not turn on when revenue dips and off when budget feels tight — consistently outperform businesses that treat it as a tap.
The tap model: run ads when things are slow, turn them off when budget feels tight, restart when revenue drops again. This creates inconsistency, destroys the algorithm's learning, damages email list health, and means you're always starting from zero.
The infrastructure model: consistent budget even in slower months. Email automations running continuously. Review generation happening every week. Creative refreshing on a schedule. The channel is always on, always learning, always building.
The infrastructure model requires more upfront trust that the investment will compound. It does.
10. Most Marketing Advice Is Written for a Different Business Than Yours
The marketing advice landscape is dominated by case studies of large brands, SaaS companies with unlimited A/B testing budgets, and agencies with armies of specialists.
Advice from those environments is often technically correct and practically useless for a 10-person ecommerce brand spending $5,000 a month on ads.
Statistical significance requires more data than you have. Best practices from a $2 million monthly budget don't translate to $20,000. Frameworks built for a 7-day decision cycle don't work for a 90-day one.
The most dangerous marketing advice is correct advice applied to the wrong context.
The only filter that works: before acting on advice, ask whether the business it came from is structurally similar to yours — same stage, same model, same economics. If it isn't, weight it accordingly.
What Actually Moves the Needle After 375 Campaigns
After all of it, the things I keep coming back to are less glamorous than most marketing content suggests.
A clear offer that solves a real problem for a specific person. Creative that earns attention rather than demanding it. Systems that run continuously so acquisition and retention don't both depend on you being present. Patience with the metrics that compound, and speed with the tests that produce learning.
That's it. Not the newest platform. Not the most sophisticated automation stack. Not the perfect campaign structure. The fundamentals, executed consistently.
The interesting thing about working in this field is that the fundamentals stay true even as everything else changes. The tactics shift. The platforms evolve. The channels that dominate one year become background noise two years later.
But the businesses that win are still, reliably, the ones that know what they're selling, who they're selling it to, and why someone would choose them. Everything else is execution.
Frequently Asked Questions About Marketing Strategy for Founders
Why don't my Meta Ads campaigns work even when I follow best practices? Most campaigns that fail weren't given the conditions to succeed. Check: was the budget high enough to exit the learning phase? Was creative changed before getting meaningful data? Was the landing page actually converting? The problem is usually one of these, not the campaign itself.
Should I focus on customer acquisition or customer retention? Both — but most founders underinvest in retention relative to acquisition. Acquiring a new customer costs 5-7x more than retaining one. Build your email automations and post-purchase sequences before you scale acquisition spend. Retention infrastructure makes acquisition economics work.
What marketing metrics actually matter for an ecommerce brand? Track MER (total revenue / total ad spend across all channels), conversion rate on your product pages, email revenue per subscriber, average order value, and customer LTV. Platform ROAS is useful for comparing creative within a channel but shouldn't drive budget decisions.
How long does it take for performance marketing to actually work? Most campaigns need 7-14 days to exit the learning phase, and another 2-4 weeks to generate enough data for meaningful optimisation. Expect 4-6 weeks before you have a clear read on whether a channel works. Founders who pull the plug at two weeks are making decisions on noise, not signal.
More to Discover
What 375 Campaigns Taught Me About Marketing
375 campaigns. 50+ clients. Some made hundreds of thousands. Some wasted money. Here are the marketing lessons that only come from doing it — not studying it.
Insights
I've run over 375 marketing campaigns across more than 50 clients. Ecommerce brands, service businesses, startups. Founders spending their last $3,000 on ads. Founders scaling through seven figures.
Some campaigns made clients hundreds of thousands. Some wasted money. Most landed somewhere in between.
I've spent a lot of time thinking about what actually separated the ones that worked from the ones that didn't. Not just tactically — which bidding strategy, which creative format — but at the level of thinking and decisions.
The lessons below aren't from courses or frameworks. They're from doing the same thing many times and watching what happened. Some are counterintuitive. Most took multiple expensive failures to actually understand.
1. The Dashboard Is Never the Full Picture
Every platform reports in its own favour.
Meta's ROAS is overstated because it claims credit for sales that Google, email, and organic also claim. Google's conversion tracking fires on sessions that were already going to convert. TikTok counts views as engagements.
None of these platforms are deliberately deceiving you. They're measuring what they can see, through their own lens. But if you make budget decisions based on any single platform's self-reported numbers, you're working with a biased and incomplete picture.
The discipline I use on every client account: pull real numbers first. Total revenue from Shopify or the CRM. Total spend across all channels. Calculate MER. Then look at individual platform data to understand the composition — not the total.
It's slower than reading a ROAS number. It's also accurate.
2. The Things That Feel Like Marketing Often Aren't
Founders spend a lot of time on things that feel like marketing but don't move revenue.
Redesigning the logo. Rewriting the website copy for the fourth time. Debating brand colour palettes. Building elaborate campaign structures before basic messaging has been proven. Creating content for an existing audience of 600 followers that won't reach anyone new.
There's a version of brand work that's genuinely productive. There's also a version that's creative procrastination with a respectable cover story.
The test I find useful: will this activity, in the next 90 days, lead to someone who doesn't know the brand finding out about it — or someone who knows it being more likely to buy? If the honest answer is no, it's probably not marketing. It might be important for other reasons. But it's not marketing.
3. Audience Quality Matters More Than Audience Size
When founders first look at Meta Ads targeting, they're often drawn to the biggest possible audiences. Millions of people. Maximum reach.
The audience that converts best in most campaigns I've run wasn't the biggest. It was the most specific — the tightest match between the product, the problem it solves, and the person experiencing that problem right now.
A cold audience of 50,000 people who match your best customer profile will almost always outperform a broad audience of 5 million, especially early on when the algorithm doesn't yet have enough conversion data to find your buyer inside a huge pool.
The shift toward broad audiences Meta encourages makes sense at scale — when you have thousands of conversions and the algorithm genuinely knows your buyer. Before you have that data, tighter targeting is usually better.
4. Most Campaigns That "Don't Work" Were Never Given the Conditions to Work
I've turned around dozens of campaigns clients told me weren't working. More often than not, the campaigns weren't the problem.
The budget was too low to exit the learning phase. Creative was swapped before any statistical significance. The campaign was paused and restarted multiple times, resetting the algorithm each time. The product page was converting at 0.4% and no amount of ad work was going to fix that.
Good campaigns fail in bad conditions.
When a campaign doesn't perform, the right diagnostic question isn't "what's wrong with this campaign?" It's "have we actually given this campaign what it needs to work?" Those are different questions with different answers. The first leads to endless tinkering. The second leads to fixing the actual constraint.
5. Retention Is Worth More Than Acquisition, but Almost Nobody Acts Like It
The maths of customer retention are well known: acquiring a new customer costs 5-7x more than retaining an existing one. Repeat customers spend more per transaction. They refer more people. LTV, not CPA, determines whether a business model is healthy.
And yet — in the vast majority of new client accounts I've worked on, the email automations are either absent or basic. Post-purchase sequences are missing. Win-back campaigns have never been built. Loyalty mechanics don't exist.
Everyone knows retention matters. Almost nobody has built the systems to act on it.
The reason is that retention work is slower and less immediately satisfying than acquisition work. Running a new Meta campaign has a feedback loop of days. Building a post-purchase email sequence has a feedback loop of months. Founders chase the quick signal.
The clients who've seen the most durable growth are the ones who invested in retention infrastructure alongside acquisition channels. The acquisition brought customers in. The retention system made the unit economics work.
6. The Offer Is Usually the Real Bottleneck
When campaigns underperform, the first instinct is to change the creative or the targeting. These are the visible variables — the things you can directly control in Ads Manager.
But when a brand has a fundamentally weak offer, no amount of creative or targeting work will rescue it.
The offer is: what are you selling, at what price, with what value proposition, against what alternatives? If the price is too high relative to perceived value, if the guarantee isn't there when competitors offer one, if the product page raises more questions than it answers — campaigns will struggle regardless of how good the ads are.
Before I optimise creative, I look at the offer. Before I change targeting, I look at the landing page conversion rate. The bottleneck is almost never where you first go looking for it.
7. Data Shows You What Happened. It Doesn't Explain Why.
Analytics is the most overrated skill in marketing and simultaneously the most underrated.
Overrated because people treat data as a substitute for thinking. A dashboard full of numbers doesn't explain itself. CTR went down — why? Conversion rate improved — because of what? Attributing causality to a metric movement is a skill, not a mechanical readout.
Underrated because most businesses don't collect or structure data well enough to draw any conclusions at all. UTM parameters inconsistent. Conversion events misfiring. GA4 set up but never reviewed.
The discipline is this: clean data infrastructure first. Then build the habit of looking at trends over time. Then — separately — develop the judgment to understand what the data is actually telling you versus what you want it to say.
Numbers show what happened. Figuring out why requires thinking, not more data.
8. Speed of Learning Beats Perfection of Execution
In the early stages of any channel or campaign, the goal isn't to launch a perfect campaign. It's to find out what works as fast as possible.
A good-enough campaign that launches in one week, generates data, and informs a better second version the following week will outperform a perfect campaign that takes six weeks to build. Learning compounds. Delay doesn't.
This is the most important mindset shift for founders who come from non-marketing backgrounds. The engineering instinct — build it right before you ship — will actively slow you down in marketing, where the right answer can only be discovered by running something and watching what happens.
Ship a testable version. Learn. Improve. Repeat. The version you're on in three months will be far better than whatever you'd have built in the first six weeks of planning.
9. The Clients Who Treat Marketing as Infrastructure Win
The most consistent pattern I've seen across 50+ clients: businesses that treat marketing as infrastructure — something to build properly and maintain, not turn on when revenue dips and off when budget feels tight — consistently outperform businesses that treat it as a tap.
The tap model: run ads when things are slow, turn them off when budget feels tight, restart when revenue drops again. This creates inconsistency, destroys the algorithm's learning, damages email list health, and means you're always starting from zero.
The infrastructure model: consistent budget even in slower months. Email automations running continuously. Review generation happening every week. Creative refreshing on a schedule. The channel is always on, always learning, always building.
The infrastructure model requires more upfront trust that the investment will compound. It does.
10. Most Marketing Advice Is Written for a Different Business Than Yours
The marketing advice landscape is dominated by case studies of large brands, SaaS companies with unlimited A/B testing budgets, and agencies with armies of specialists.
Advice from those environments is often technically correct and practically useless for a 10-person ecommerce brand spending $5,000 a month on ads.
Statistical significance requires more data than you have. Best practices from a $2 million monthly budget don't translate to $20,000. Frameworks built for a 7-day decision cycle don't work for a 90-day one.
The most dangerous marketing advice is correct advice applied to the wrong context.
The only filter that works: before acting on advice, ask whether the business it came from is structurally similar to yours — same stage, same model, same economics. If it isn't, weight it accordingly.
What Actually Moves the Needle After 375 Campaigns
After all of it, the things I keep coming back to are less glamorous than most marketing content suggests.
A clear offer that solves a real problem for a specific person. Creative that earns attention rather than demanding it. Systems that run continuously so acquisition and retention don't both depend on you being present. Patience with the metrics that compound, and speed with the tests that produce learning.
That's it. Not the newest platform. Not the most sophisticated automation stack. Not the perfect campaign structure. The fundamentals, executed consistently.
The interesting thing about working in this field is that the fundamentals stay true even as everything else changes. The tactics shift. The platforms evolve. The channels that dominate one year become background noise two years later.
But the businesses that win are still, reliably, the ones that know what they're selling, who they're selling it to, and why someone would choose them. Everything else is execution.
Frequently Asked Questions About Marketing Strategy for Founders
Why don't my Meta Ads campaigns work even when I follow best practices? Most campaigns that fail weren't given the conditions to succeed. Check: was the budget high enough to exit the learning phase? Was creative changed before getting meaningful data? Was the landing page actually converting? The problem is usually one of these, not the campaign itself.
Should I focus on customer acquisition or customer retention? Both — but most founders underinvest in retention relative to acquisition. Acquiring a new customer costs 5-7x more than retaining one. Build your email automations and post-purchase sequences before you scale acquisition spend. Retention infrastructure makes acquisition economics work.
What marketing metrics actually matter for an ecommerce brand? Track MER (total revenue / total ad spend across all channels), conversion rate on your product pages, email revenue per subscriber, average order value, and customer LTV. Platform ROAS is useful for comparing creative within a channel but shouldn't drive budget decisions.
How long does it take for performance marketing to actually work? Most campaigns need 7-14 days to exit the learning phase, and another 2-4 weeks to generate enough data for meaningful optimisation. Expect 4-6 weeks before you have a clear read on whether a channel works. Founders who pull the plug at two weeks are making decisions on noise, not signal.
More to Discover
What 375 Campaigns Taught Me About Marketing
375 campaigns. 50+ clients. Some made hundreds of thousands. Some wasted money. Here are the marketing lessons that only come from doing it — not studying it.
Insights
I've run over 375 marketing campaigns across more than 50 clients. Ecommerce brands, service businesses, startups. Founders spending their last $3,000 on ads. Founders scaling through seven figures.
Some campaigns made clients hundreds of thousands. Some wasted money. Most landed somewhere in between.
I've spent a lot of time thinking about what actually separated the ones that worked from the ones that didn't. Not just tactically — which bidding strategy, which creative format — but at the level of thinking and decisions.
The lessons below aren't from courses or frameworks. They're from doing the same thing many times and watching what happened. Some are counterintuitive. Most took multiple expensive failures to actually understand.
1. The Dashboard Is Never the Full Picture
Every platform reports in its own favour.
Meta's ROAS is overstated because it claims credit for sales that Google, email, and organic also claim. Google's conversion tracking fires on sessions that were already going to convert. TikTok counts views as engagements.
None of these platforms are deliberately deceiving you. They're measuring what they can see, through their own lens. But if you make budget decisions based on any single platform's self-reported numbers, you're working with a biased and incomplete picture.
The discipline I use on every client account: pull real numbers first. Total revenue from Shopify or the CRM. Total spend across all channels. Calculate MER. Then look at individual platform data to understand the composition — not the total.
It's slower than reading a ROAS number. It's also accurate.
2. The Things That Feel Like Marketing Often Aren't
Founders spend a lot of time on things that feel like marketing but don't move revenue.
Redesigning the logo. Rewriting the website copy for the fourth time. Debating brand colour palettes. Building elaborate campaign structures before basic messaging has been proven. Creating content for an existing audience of 600 followers that won't reach anyone new.
There's a version of brand work that's genuinely productive. There's also a version that's creative procrastination with a respectable cover story.
The test I find useful: will this activity, in the next 90 days, lead to someone who doesn't know the brand finding out about it — or someone who knows it being more likely to buy? If the honest answer is no, it's probably not marketing. It might be important for other reasons. But it's not marketing.
3. Audience Quality Matters More Than Audience Size
When founders first look at Meta Ads targeting, they're often drawn to the biggest possible audiences. Millions of people. Maximum reach.
The audience that converts best in most campaigns I've run wasn't the biggest. It was the most specific — the tightest match between the product, the problem it solves, and the person experiencing that problem right now.
A cold audience of 50,000 people who match your best customer profile will almost always outperform a broad audience of 5 million, especially early on when the algorithm doesn't yet have enough conversion data to find your buyer inside a huge pool.
The shift toward broad audiences Meta encourages makes sense at scale — when you have thousands of conversions and the algorithm genuinely knows your buyer. Before you have that data, tighter targeting is usually better.
4. Most Campaigns That "Don't Work" Were Never Given the Conditions to Work
I've turned around dozens of campaigns clients told me weren't working. More often than not, the campaigns weren't the problem.
The budget was too low to exit the learning phase. Creative was swapped before any statistical significance. The campaign was paused and restarted multiple times, resetting the algorithm each time. The product page was converting at 0.4% and no amount of ad work was going to fix that.
Good campaigns fail in bad conditions.
When a campaign doesn't perform, the right diagnostic question isn't "what's wrong with this campaign?" It's "have we actually given this campaign what it needs to work?" Those are different questions with different answers. The first leads to endless tinkering. The second leads to fixing the actual constraint.
5. Retention Is Worth More Than Acquisition, but Almost Nobody Acts Like It
The maths of customer retention are well known: acquiring a new customer costs 5-7x more than retaining an existing one. Repeat customers spend more per transaction. They refer more people. LTV, not CPA, determines whether a business model is healthy.
And yet — in the vast majority of new client accounts I've worked on, the email automations are either absent or basic. Post-purchase sequences are missing. Win-back campaigns have never been built. Loyalty mechanics don't exist.
Everyone knows retention matters. Almost nobody has built the systems to act on it.
The reason is that retention work is slower and less immediately satisfying than acquisition work. Running a new Meta campaign has a feedback loop of days. Building a post-purchase email sequence has a feedback loop of months. Founders chase the quick signal.
The clients who've seen the most durable growth are the ones who invested in retention infrastructure alongside acquisition channels. The acquisition brought customers in. The retention system made the unit economics work.
6. The Offer Is Usually the Real Bottleneck
When campaigns underperform, the first instinct is to change the creative or the targeting. These are the visible variables — the things you can directly control in Ads Manager.
But when a brand has a fundamentally weak offer, no amount of creative or targeting work will rescue it.
The offer is: what are you selling, at what price, with what value proposition, against what alternatives? If the price is too high relative to perceived value, if the guarantee isn't there when competitors offer one, if the product page raises more questions than it answers — campaigns will struggle regardless of how good the ads are.
Before I optimise creative, I look at the offer. Before I change targeting, I look at the landing page conversion rate. The bottleneck is almost never where you first go looking for it.
7. Data Shows You What Happened. It Doesn't Explain Why.
Analytics is the most overrated skill in marketing and simultaneously the most underrated.
Overrated because people treat data as a substitute for thinking. A dashboard full of numbers doesn't explain itself. CTR went down — why? Conversion rate improved — because of what? Attributing causality to a metric movement is a skill, not a mechanical readout.
Underrated because most businesses don't collect or structure data well enough to draw any conclusions at all. UTM parameters inconsistent. Conversion events misfiring. GA4 set up but never reviewed.
The discipline is this: clean data infrastructure first. Then build the habit of looking at trends over time. Then — separately — develop the judgment to understand what the data is actually telling you versus what you want it to say.
Numbers show what happened. Figuring out why requires thinking, not more data.
8. Speed of Learning Beats Perfection of Execution
In the early stages of any channel or campaign, the goal isn't to launch a perfect campaign. It's to find out what works as fast as possible.
A good-enough campaign that launches in one week, generates data, and informs a better second version the following week will outperform a perfect campaign that takes six weeks to build. Learning compounds. Delay doesn't.
This is the most important mindset shift for founders who come from non-marketing backgrounds. The engineering instinct — build it right before you ship — will actively slow you down in marketing, where the right answer can only be discovered by running something and watching what happens.
Ship a testable version. Learn. Improve. Repeat. The version you're on in three months will be far better than whatever you'd have built in the first six weeks of planning.
9. The Clients Who Treat Marketing as Infrastructure Win
The most consistent pattern I've seen across 50+ clients: businesses that treat marketing as infrastructure — something to build properly and maintain, not turn on when revenue dips and off when budget feels tight — consistently outperform businesses that treat it as a tap.
The tap model: run ads when things are slow, turn them off when budget feels tight, restart when revenue drops again. This creates inconsistency, destroys the algorithm's learning, damages email list health, and means you're always starting from zero.
The infrastructure model: consistent budget even in slower months. Email automations running continuously. Review generation happening every week. Creative refreshing on a schedule. The channel is always on, always learning, always building.
The infrastructure model requires more upfront trust that the investment will compound. It does.
10. Most Marketing Advice Is Written for a Different Business Than Yours
The marketing advice landscape is dominated by case studies of large brands, SaaS companies with unlimited A/B testing budgets, and agencies with armies of specialists.
Advice from those environments is often technically correct and practically useless for a 10-person ecommerce brand spending $5,000 a month on ads.
Statistical significance requires more data than you have. Best practices from a $2 million monthly budget don't translate to $20,000. Frameworks built for a 7-day decision cycle don't work for a 90-day one.
The most dangerous marketing advice is correct advice applied to the wrong context.
The only filter that works: before acting on advice, ask whether the business it came from is structurally similar to yours — same stage, same model, same economics. If it isn't, weight it accordingly.
What Actually Moves the Needle After 375 Campaigns
After all of it, the things I keep coming back to are less glamorous than most marketing content suggests.
A clear offer that solves a real problem for a specific person. Creative that earns attention rather than demanding it. Systems that run continuously so acquisition and retention don't both depend on you being present. Patience with the metrics that compound, and speed with the tests that produce learning.
That's it. Not the newest platform. Not the most sophisticated automation stack. Not the perfect campaign structure. The fundamentals, executed consistently.
The interesting thing about working in this field is that the fundamentals stay true even as everything else changes. The tactics shift. The platforms evolve. The channels that dominate one year become background noise two years later.
But the businesses that win are still, reliably, the ones that know what they're selling, who they're selling it to, and why someone would choose them. Everything else is execution.
Frequently Asked Questions About Marketing Strategy for Founders
Why don't my Meta Ads campaigns work even when I follow best practices? Most campaigns that fail weren't given the conditions to succeed. Check: was the budget high enough to exit the learning phase? Was creative changed before getting meaningful data? Was the landing page actually converting? The problem is usually one of these, not the campaign itself.
Should I focus on customer acquisition or customer retention? Both — but most founders underinvest in retention relative to acquisition. Acquiring a new customer costs 5-7x more than retaining one. Build your email automations and post-purchase sequences before you scale acquisition spend. Retention infrastructure makes acquisition economics work.
What marketing metrics actually matter for an ecommerce brand? Track MER (total revenue / total ad spend across all channels), conversion rate on your product pages, email revenue per subscriber, average order value, and customer LTV. Platform ROAS is useful for comparing creative within a channel but shouldn't drive budget decisions.
How long does it take for performance marketing to actually work? Most campaigns need 7-14 days to exit the learning phase, and another 2-4 weeks to generate enough data for meaningful optimisation. Expect 4-6 weeks before you have a clear read on whether a channel works. Founders who pull the plug at two weeks are making decisions on noise, not signal.