Learn marketing from people who do it

Positioning is context, not copy 03:02

The same product means something different depending on the category you put it in: the category sets the competitors, the price and what buyers assume it does.

It is double chocolate caramel cake, but if I put it in the muffin format and I call it a muffin, everything about that is different.

Win a small segment the leader ignores, then expand 13:16

A startup can't beat the market leader head-on, so it picks a subsegment where its one uncopyable feature matters most, dominates it, and widens from there.

we narrowed down the positioning instead of saying we were enterprise CRM, we were CRM for investment banks

In B2B the emotion is fear, and the real rival is no decision 20:56

Business buyers are afraid of picking wrong in front of their boss, which favours the safe incumbent and makes many deals end with nobody choosing anything.

In B2B, the granddaddy of emotions is fear.

Target first, research second 17:55

Without money for big surveys, build a rough map of the market, pick where to play, then go and spend time with eight or ten of those customers to work out positioning.

rather than doing research early on to build the map of the market we're actually going to do what I would say is targeted ethnography

Position against the big guy 20:30

A small brand can define itself against a famous rival, borrowing its fame and finding the gap in how it plays; the big brand has no reason to answer.

you can also position against an alternative more Famous Brand in the market

Strategy is what you don't do 26:42

Strategy means choosing who you won't serve, what you won't stand for and what you won't chase; that matters more for a small company than a big one.

remember strategy is best described as what we do not do

Brand building and activation work on different clocks 22:38

Activation (search ads, offers) catches people buying now and pays back fast; brand work plants memories with the many people who will buy later. You need both.

most important word in the long and the short of it is and

Brand ads make people feel, not know 31:22

Short-term ads inform and push a sale; long-term brand ads work through emotion, music and craft, and can't be reduced to a message.

It's not really about telling people things. It's about making people feel things.

Ads wear out far slower than marketers think 41:39

Teams get bored of a campaign long before most of the public has seen it; most ads are replaced before they stop working.

most ads never get a chance to wear out at all

Mental availability is about cues, not awareness 17:51

Everyone knows Coca-Cola, yet it rarely comes to mind at the moment someone picks a drink; the job is linking the brand to the situations where people buy.

mental availability is a share thing of like, well, how often do they?

Advertising keeps a brand up more than it lifts it 11:48

Ads are a weak force: doubling a budget changes little, but stopping slowly loses ground. Big brands often got big through history and luck, and stay big through marketing.

it's better at keeping the plane up when it's already up than it is of driving the plane up there in the first place

Distinctive, not differentiated 37:25

Brand assets like Geico's gecko or Yellow Tail's kangaroo help people recognise and find you; they are not reasons to buy, and competing brands are rarely as different as marketers think.

They're just looking like you so that when when a consumer when there is a touch point, they realize which brand it is.

Change the perception instead of the product 08:11

Many problems treated as engineering problems are really psychology problems, and fixing how something feels is often far cheaper than fixing what it is.

If it's cheaper to reduce anxiety than to increase range, that's what we should do.

Value is relative 15:57

People judge a price against what sits next to it, so showing economy and premium side by side sold more premium seats than making people pick a class first.

It's a relative measure not an absolute measure.

Reverse benchmarking 33:50

Don't copy what competitors do well; find what the whole category does badly and do that brilliantly, because surprise is what people remember.

Find out what the other people are doing badly and do it really, really well.

Category entry points are about buyers, not brands 21:02

Map the situations and needs people have before any brand comes to mind (packing the kids' lunch, needing coffee after a meeting), then link your brand to the common ones. 'Trust' is not one, because it describes a brand.

So they're not about brands, they're about buyers.

A brand asset needs fame and uniqueness 08:18

A logo, colour or character only works if most people link it to you (fame) and to nobody else (uniqueness); a shared one spends your money reminding people of competitors.

Uniqueness is about do we own this asset? Are we the only brand that comes to mind when this asset is present?

Light buyers carry the brand 28:43

Most of a brand's buyers buy it rarely and hold only faint memories of it, so they are the first to slip away when marketing stops. Reaching them is growth and insurance at once.

when brands decline, it's light buyers that drop off first

Measure lift, not clicks 03:14

Coca-Cola could not track a single billboard viewer, so it compared sales in a city with the billboard against a similar city without it. You can still measure marketing that way.

we think that this billboard probably provided 1.5% lift in the geographic area where people saw it

The flyer kid in the alley 26:40

Ad platforms often show ads to people who were already on their way to buy, then take credit for the sale. A channel that looks great in attribution reports may not be creating any new demand.

why would you work hard to make more sales when you can just take credit for the sales that were already going to happen

Brand searches come from somewhere else 35:31

When someone hears about you and then Googles your name, Google gets the credit for the click. The demand was created elsewhere, so don't give search the credit for it.

no search for a brand deserves to be attributed to Google that demand was created somewhere else

Value before the click 02:03

Give people the complete useful thing where they already are, rather than holding back the payoff to lure them to your site. People will look you up when they're ready.

It means creating value before the click and not being dependent on that click to happen every single time.

Take your idea on a tour 12:26

Test an idea as a short post, run it as a webinar where people push back, then turn it into a blog post, podcast talking points and finally a talk. Each round sharpens the idea.

the webinar would be it would basically be like my draft blog post

Search captures demand, it doesn't create it 46:40

People only search for what they already know about. What gets said about you in public is what makes them want it in the first place.

search only captures demand, but the public record creates it

Retrieved, cited, trusted 04:09

Being crawled by an AI doesn't get you cited, and being cited doesn't get you picked. In his study, users chose the top item on an AI shortlist unless a brand they already trusted appeared on it.

The only exception to that is when they see a brand on their short list that they trust.

Optimize for the bot's searches 07:48

AI tools run their own background searches before answering. The queries that matter may have no human search volume at all.

we have to optimize more for keywords that might not have search volume, at least not from humans, but from bots.

Join a conversation people already care about 18:40

Original data gets cited, but only when it answers a question your buyers are already asking. Look at what is happening in your market first, then find data that speaks to it. Don't start from whatever data you happen to have.

it's only going to get attention if you join a conversation that is relevant to the people you're trying to talk to

Visualize, falsify, nobody else can say it 00:00

Check every line with three questions: can I picture it, could it be proved true or false, and could only we say it? Three no's mean filler.

Can I visualize it? Can I falsify it? Can nobody else say this?

Don't talk, point 08:16

Rather than describing something with adjectives, point at evidence: a chart, a fact, a person. It makes you do the research.

It it it basically it gets you off the adjective trail.

From point A to point B 21:39

Before writing, name what the reader thinks now and what you want them to think. Then work out who you're talking to, what you have to say and how to say it well.

what's the current attitude? What's the desired attitude? That's point A, that's point B.

Copy is 90% listening 03:09

Only a small part of the job is writing. The rest is collecting the words customers actually use, because your own opinion on why someone should buy doesn't count.

is just listening to customers because nobody cares about the English majors opinion on why you should buy that piece of software

Breakthrough or bust 09:27

Before testing a line, ask whether it's different enough that it could either win big or fail badly. A safe line won't teach you anything.

ask our copy is this pushing things far enough that could either be a dramatic breakthrough or a total bust

Documentary-style moments 22:29

In sales calls, listen for moments so vivid someone could act them out. Nobody in a meeting room would think of them, and they make copy specific.

could they act it out and that's a really good moment that's something that you can't come up with in a boardroom

Talk onto the page 00:32

Weak writers compose word by word, while good ones write the way they'd persuade someone in person. Everyone can already persuade; the trick is using that skill when you sit down to write.

the best writers it's like they're just talking onto the page

The magic copy square 20:48

Plot benefit against believability. Big promise with no proof is hype, proof with no promise is 'who cares', and you need both to sell.

if you have enough benefit and you've got a lot of believability then it's like sold

Ask 'so what' until it matters 24:30

Take a benefit and keep asking what it means for the reader's life. 'Less arthritis pain' becomes 'you can pick up your grandkids'.

you'll make more money well so what what does that mean

Growth model, not business model 0:08:00

Before picking tactics or metrics, write down how one new user leads to the next one. A financial model is about money in and out; a growth model is about users in and out, and its levers are different.

When I put a user in, how do I get more than one user out?

Product-channel fit 0:22:30

Platforms like Google, Facebook and Apple set the rules, so the product has to be shaped to work inside one channel, not the other way round. Building first and bolting on distribution later usually fails.

We cannot mold channels to products. We have to mold the product to the channel

Channel-model fit 0:23:30

Each channel only works at certain price points. Virality and paid ads suit cheap, low-friction products; sales suits expensive ones. Price and channel have to match.

So you can't use viral loops for a product that you're charging 10 thousand dollars up front for.

Atomic network 0:13:30

Work out the smallest group that has to be on the product at once for it to be useful (2-3 for Zoom, a team for Slack, hundreds of drivers in a city for Uber), win that group, then copy it.

how many people need to be connected at the same time doing the same thing in order for that product to be valuable

Hard side vs easy side 0:18:30

One side of a network (creators, sellers, drivers) does more work and is pickier and more valuable. Build features that make that side happy, as Tinder's swipe did for its most sought-after users.

every network product has what I describe as an easy side and a hard side of the network

Invites copy the network 0:38:30

Invite-only launches work less because of hype than because every invited person already has a connection inside, so new users land in a group that already works.

when you are invited into a product, that means you are already connected with one other person

Distribution beats product 0:01:00

A great product with no built-in way to reach people dies slowly, while mediocre products with strong distribution win. Product teams have to own distribution, not hand it to marketing.

what builds a successful company is having a great distribution for your product

Loops over funnels 0:03:00

The fastest-growing companies grow through cycles where a user's action brings in the next user (Dropbox shared files drive 60% of its acquisition), not through a one-way funnel you keep refilling.

They don't grow via funnels. They grow via loops.

Free product as marketing budget 0:17:30

At Lovable most cost is free usage, and they count it as marketing spend: better to give the product away than to pay ad platforms for attention.

we look at it not as a cost center. We're looking it as our marketing budget

Win the individual, then let them sell it inside 0:14:00

Figma's motion was bottom-up: make the people who use the tool all day love it so much that they spend their own credibility getting their company to adopt it. No sales team for the first three years.

they love you so much that they're willing to put their social capital and themselves in the line and spread the product

Don't market to experts; earn credibility 0:34:00

Designers ignored buzzwords, so Figma published deep technical posts by its engineers and designers. Her test for whether a post was good enough: could the marketer have written it?

if I could have written it myself it was probably too basic

Go where they already are 0:44:30

Early on nobody will come to your Slack or forum. Find the one place your audience already gathers (for designers, Twitter), map who influences whom, and put your effort there only.

where can you go where they already are as opposed to making them come to you

Launch is a habit, not an event 0:01:00

Founders spend months polishing one launch, then nobody notices. Launch early and repeatedly (Airbnb launched three times) and treat each one as a test.

I want you to think about launching as something that you do continually

Say what before why 0:05:00

Open with the company name and what it does, in plain words with no jargon, so the listener could roughly picture what you built. The backstory can come later if they ask.

we recommend leading with what not with why

Each channel is a different audience 0:12:00

A landing page, friends and family, strangers, online communities, a waitlist: each launch reaches different people, so launching in many places also tells you who your users really are. Press won't do this for you.

different channels will naturally reach different audiences

People buy from people 0:04:30

In a crowded market where products look alike, buyers pick the company whose founder they already follow and trust. Early Drift customers came because they liked David Cancel's podcast and point of view.

people want to work with people they feel like they know and they like and they trust

Your story is your strategy 0:08:30

Drift positioned itself as chat 'built for sales and marketing' against Intercom's product focus, and the roadmap had to keep proving it. The message only works if the product backs it every week.

this is why people say your story is your strategy

Post often to test ideas 0:19:30

Posting most days is not about reach; each post is a test of an idea with your ideal customers, and what lands becomes the newsletter, the talk, the book. You still need two or three strong opinions to say.

the whole magic in social media is the feedback loop