142 things you should know if you work in consumer growth today
2026 edition // edition one.
Inspired by New York Mag’s recent 247 things you should know as a New Yorker, we thought we’d compile our own version for you.
On measurement
ROAS is a vanity metric in a profitability costume.
Contribution margin per session tells you more about a DTC brand’s health than ROAS ever will. CM3 = revenue minus cost of goods, minus fulfilment, minus paid media spend.
Blended ROAS conceals whether you’re winning or subsidising losses with organic.
If you’re optimising for ROAS, you are not optimising for incrementality.
MER (marketing efficiency ratio) is blended ROAS with a different name and the same limitations.
On Meta
Meta is a scale channel, not an efficiency channel.
CTR does not predict CPA.
Thumbstop rate does not predict CPA.
If you’re not running awareness spend, your conversion campaigns are doing the work of warming cold audiences and paying conversion CPMs to do it.
You should evaluate awareness investment after a minimum of three months. Judging it on last-click in week two is a category error.
The 45+ demographic is cheaper to reach and often has higher AOV and LTV. Most brands under £5m ignore it because their founders are 28.
Logos and CTAs in ads actually perform better on average, which is annoying if you built your whole brand around looking “authentic.”
Static ads frequently beat video because the message lands before the thumb moves.
The biggest mistake in DPA setup is combining incompatible product sets under one campaign.
I’ve seen accounts with 1x campaign, 1x ad set scale to £350k/month.
On creative
Jobs-to-be-done research starts not with what people buy, but with what they were doing before they searched.
The four forces of progress are: push, pull, anxiety, and habit. The most important two for paid social are anxiety and habit.
1 star reviews are more valuable than 5 star reviews.
“I’ve tried everything” is not a customer complaint. It’s a hook.
UGC can be scripted. It only fails with bad copywriting or acting.
Yappers, talking-head videos of real people just talking, are quietly the top-performing format in several DTC categories.
When talking to customers always preface anxieties with “apart from price, what....”
Your brand tagline is almost certainly your worst-performing ad headline. Taglines are written for boardrooms, not for the scroll.
Review mining at scale is most efficiently done with AI, if you give it the right taxonomy first.
Creative fatigue isn’t as simple as many let on. Turn old ads back on. Go on.
Testimonial ads work because they remove anxiety rather than add aspiration.
“You put olive oil on that?” is a great hook because it assumes knowledge the viewer is compelled to validate. Use your customers’ language.
On strategy
Full-funnel advertising lowers CPMs on conversion campaigns. It’s not just brand investment, it’s efficiency investment.
At every growth stage, ask: what’s the single constraint that, if removed, unlocks the next order of magnitude?
“Diversify away from Meta” is advice given by those who don’t understand most consumer ecom/DTC.
Be aware of LTV debt: don’t get to £10m/year with the same LTV you had at £1m.
At a certain stage, you have to accept your natural CPA. If it doesn’t work, you don’t have product-market fit.
Incrementality testing is the closest thing we have to answering “would these customers have bought anyway?”
If your gross margin is under 50%, paid acquisition is extremely difficult to make profitable after fulfilment and media costs.
Track profit per employee. It forces you to confront whether each hire is actually driving margin or just making the org chart look right.
On experiments
Most creative testing on Meta is flawed because the algorithm redistributes spend within ad sets based on early signals. You’re not testing creative in isolation, you’re testing what Meta chooses to spend on.
A creative test needs roughly 20 purchases before you can call a result with any confidence.
80% net new creative, 20% iteration.
Media buying experiments need a month, not 5 days.
Experiment velocity matters more than confidence per experiment, at early stage especially.
Statistical significance is misused in most DTC contexts because sample sizes are too small and observation windows too short.
On AI and what it actually changes
AI lets a growth team produce the creative volume of a team three times its size. The advantage isn’t cost reduction, it’s test velocity.
AI video works best when it’s either immediately obvious, or not noticeable at all.
The gap between “someone who uses AI” and “someone who’s AI-native” is whether AI is the workflow or adjacent to it.
Learn to build with AI coding tools now. The growth people who can spin up custom dashboards, scrapers, and analysis tools without waiting for engineering will have a compounding advantage for the next two to three years.
AI fluency is a temporary advantage. As tools become easier and more embedded, the differentiator will shift back to judgement, taste, and domain knowledge.
“AI-native” will be a redundant phrase by 2028, for the same reason “internet-native business” became redundant. It’ll just be what you are.
On data
A blended CPA of £24 means nothing without knowing what the customer bought and whether they came back.
Power analysis before an incrementality test tells you whether your test cells are large enough to detect the effect you actually care about.
Most “data-driven” decisions in marketing are intuition-driven with data bolted on afterwards.
Invest in your data ETL and warehouse.
On category dynamics
Health and wellness DTC brands have a structural advantage in creative because pain is a universal hook.
Visual DTC (fashion etc) leans into huge volumes of different creator/people so someone can recognise themself somewhere.
There are no hard categories, just people who ignore the rules of Meta.
If you can get your F&B work DTC alone, then your retail will fly.
The 45+ female demographic is the most underserved audience on Meta and often has the highest LTV.
Whether you’re dtc or retail-first, don’t add a second sales channel until £5m of revenue.
On the industry
The claim that “native beats brand” in creative is true on average, but not exclusively. You need a mix.
Video receives roughly 2x the budget of statics across most accounts. Statics often convert better.
AI-generated ads are a volume lever, not a creative strategy.
The reason most performance agencies don’t evolve into strategy firms is that strategy is harder to sell and slower to prove.
The biggest structural shift in DTC since 2020 has been the collapse of reliable last-click attribution as iOS changed everything.
On growth stage
From £0 to £1m, the constraint is almost always creative and targeting, finding what works.
From £1m to £5m, the constraint shifts to operational capacity and repeatable testing processes.
From £5m to £10m, the bottleneck is usually unit economics and whether the model holds at scale.
The moment to introduce a full-funnel strategy is when you can see CPMs rising on conversion campaigns without a corresponding CPA improvement.
A brand that can’t survive a three-month Meta creative drought doesn’t have a sustainable acquisition model. It has a hot streak.
Email and SMS matter most in the £1m to £5m revenue band, where repeat purchase rate determines whether the model is profitable.
On psychology and persuasion
Social proof works differently at the top and bottom of funnel. At the top you’re reducing scepticism. At the bottom you’re removing the last reason not to buy.
Fear-based creative underperforms in beauty and lifestyle categories where aspiration is the primary purchase driver.
Anxiety (one of the four forces of progress) is reduced most effectively in creative by social proof and specificity, not reassurance statements.
The “exhaustion arc” customer story works because it validates every previous failed attempt before offering a solution.
Reviews from the 45+ demographic convert better because they tend to be specific, context-rich, and written by people who’ve already tried everything else.
On what most people get wrong
Most DTC founders underinvest in creative volume and overinvest in creative polish.
Most performance agencies optimise for reported ROAS because it’s easier to present than contribution margin.
Most creative testing is actually budget testing masquerading as creative testing.
Most retention strategies are just email sequences with a better subject line. Not a repeat purchase model.
Most people who claim to understand LTV don’t understand the margin at which that LTV was generated.
Most agency pitches fail because they answer “why us” instead of “why now, and why does this matter to you specifically.”
On what almost nobody knows
Cost per reach going up while CVR also goes up is not a contradiction. It means Meta is spending into higher-value audiences.
A brand with a 3.5x ROAS might be losing money on every customer if COGS, fulfilment, and variable costs are high enough.
The incrementality of branded search is lower than most brands assume. Many of those clicks would have happened anyway. Assume it’s 0.1x.
“Meta’s algorithm has gotten smarter” usually means “we got better at brief writing and structure.” The credit rarely belongs entirely to the platform.
The measurement problem that MMM still can’t solve well is real-time responsiveness. It’s a retrospective tool used as a planning tool.
On writing and content
The most shared pieces of content make the reader feel smart, not informed, smart.
Contrarian takes work best when the conventional wisdom you’re attacking is genuinely prevalent, not a straw man.
The difference between Grade 7 writing and Grade 12 writing is that Grade 7 says what it means on the first read.
The best hook for a founder-vulnerability post opens in the middle of the failure, not with context.
The best Substack growth lever is not frequency. It’s having a clear reason to exist that nobody else fills.
On how growth teams work in 2026
Every growth team is now somewhere on a spectrum from “we use AI occasionally” to “we wouldn’t function without it.” The gap is widening fast.
The growth team that treats creative as a hypothesis rather than a deliverable will outpace the one that treats it as a production output.
“Move fast and break things” was a software company mantra applied to marketing by people who didn’t understand the difference. You can’t A/B test a brand perception.
The thing most growth teams under-budget for is time to think. The calendar fills with execution and the strategy atrophies.
On how consumer brands work in 2026
The brands that tried to be “authentic” by mimicking UGC without the actual user part have mostly been found out.
Founder-led content performs better than brand content for most sub-£20m DTC companies. Not because it’s personal, but because it’s legible, you know who’s speaking.
Most brands building “communities” are actually building email lists with extra steps.
The Dyson problem: great product, brutal creative. The reverse exists too. Creative that oversells a product is worse than creative that undersells it, because the reviews will eventually kill you.
The brands winning on TikTok in 2026 are mostly winning because someone in the building is genuinely interested in making content, not because they cracked an algorithm.
On what it actually feels like to work in growth right now
The feeling of having a great creative test result land on a Friday is still one of the best things about this job. It hasn’t been automated yet.
The worst thing about performance marketing culture is the tendency to confuse busyness with rigour.
Lean is the new default, and that’s going to require a mindset shift for psychological comfort.
The most underrated skill in growth is writing. Not copywriting. Writing, the ability to explain what you’re doing and why to someone who isn’t in the room.
The growth meeting that never ends is the one where nobody has established what decision needs to be made by the end of it.
Context switching is the new norm, and AI only accelerates that.
On the agency relationship
The best agencies act like lawyers, not yes people
At the moment AI hasn’t helped with cost.
Your agency should challenge your thinking all the time.
Agencies get you access to usually over £1m of head count per account. That’s irreplaceable in-house.
On the craft
Writing a brief forces you to know whether you actually understand the customer. If you can’t write it, you don’t.
Best practice” in performance marketing has a shorter half-life than most people admit. What worked in 2025 is often actively wrong in 2026.
Spending two hours doing customer interviews is usually worth more than any AI assisted research.
On things that are genuinely changing in 2026
Creative production that used to take a week now takes a day. The time saved is not being reinvested in thinking, in most teams.
Incrementality is finally getting the attention it always deserved.
The post-cookie world turned out to be fine for everyone who wasn’t entirely dependent on third-party data. That group was larger than expected.
Growth people are becoming engineers.
The brands winning on YouTube in 2026 figured out that long attention and short attention require fundamentally different creative approaches, and stopped applying the same creative to both.
Subscription models in DTC turned out to be harder than they looked when every brand launched one in 2020-2022. The ones that work have genuine habitual use cases.
Some brands are realising they have audiences, not communities, and that’s fine.
On things that haven’t changed
Meta advertising is advertising.
Advertising has to sell.
Simplicity beats complexity.
Operating at 100% capacity creates a fragile business, build in space to crack.
On the wider world that growth operates in
The cost of living crisis changed consumer behaviour more than most teams have updated their models to reflect.
The brands winning in grocery and everyday essentials are the ones that made the implicit trade-off explicit: this is cheaper, or this is better, pick one.
The environmental claims conversation has become complicated enough that most brands have quietly retreated from it. The ones with genuine credentials now have more white space.
Gen Z’s relationship with brands is not what most brand strategists predicted. They’re not inherently anti-brand. They’re just better at detecting nonsense.
The resurgence of in-person retail is real and most DTC brands still treat it as secondary. The ones blending both well are building the most defensible positions.
Social commerce is still not quite the thing it kept being predicted to be in the UK. The infrastructure isn’t there yet.
The creator economy has matured. The hobbyists have mostly dropped out. What’s left is more professional, more expensive, and more effective when briefed well.
Economic uncertainty has lengthened consumer consideration cycles in discretionary categories. Most brands haven’t updated their channel mix to reflect this.
On the culture around the job
Most growth Twitter/X is people restating what worked for them as universal law. Read it, but hold it loosely.
“Founder mode” became a phrase that was used to justify almost anything. The useful version of it is simple: stay close to the work.
The tendency of growth people to dress up industry gossip as market insight is a known failure mode. Name the pattern so you can catch yourself doing it.
And finally
The best DTC brand you know is probably one you heard about from a friend, not an ad. That’s still the industry’s most important unsolved problem.
The teams that ran the best experiments in the last year are the ones that wrote down what they expected to happen before they ran them.
The best briefs are written in the customer’s language, not the strategist’s.
Every successful DTC brand has a moment when a single creative unlocked scale. The lesson is not to try to recreate that creative. It’s to build the system that finds it again.
The teams most resistant to AI adoption are often the most skilled at their current way of working. Capability and change readiness are different things. This is the individual’s version of the Innovator’s Dilemma.
The growth industry in 2026 is better at measurement, better at structure, and still not great at the thing that actually drives results: making things people find interesting.
The question a founder should ask before hiring any performance agency: “Can you show me a client where you improved margin, not just ROAS?”
The question that reveals more about a brand’s health than any dashboard: “What’s your repeat purchase rate at 90 days?”
The 90-day repeat purchase rate question at the end is the right one to end on. Most brands can tell you their overall repeat rate but very few can tell you what it looks like at 90 days by acquisition channel, and that breakdown is usually where the real insight lives.