Before joining Sydecar, my marketing experience came down to a college internship and two years post-grad as the marketing coordinator at a children’s museum, where I was focused on driving memberships, managing our social media accounts, and booking birthday parties. I spent the next four years after that job doing revenue operations, where I worked closely with sales and support teams, rarely interacting with marketing.  

And yet here I am, five-plus years into a career as a marketing leader, wondering how I got here. I often find myself feeling out of the loop in conversations with peers, or even with my direct reports. It seems like they all read the same book, took the same course, learned the same best practices for B2B marketing. I have been known to Google things like “how to calculate CAC” or “what is the difference between SEO and SEM.”

But, outside of the unavoidable bouts of imposter syndrome, I’m mostly incredibly grateful for my lack of traditional marketing experience. Not only has it made me scrappy (see: Googling how to do my job), but it’s also allowed me to enter every job with a fresh set of eyes, unencumbered by the “way things have always been done,” and instead always seeking to discover how differently things could be done. I now strongly believe that the next generation of great marketers will have no marketing background at all. 

I came to my role at Sydecar through a fascination with the venture capital industry. I had entered the industry, somewhat accidentally, just a year before knowing close to nothing. Lucky for me, it was mid-2020, and there was not much to do besides hang out on the internet. I learned a lot. Including the fact that venture capital, like so much of financial services, operates in an incredibly antiquated manner. I couldn’t believe that the people tasked with finding the next generation of world-changing companies were running their business out of spreadsheets and email threads.

This was especially shocking to me coming from revenue operations, where everything centered on optimization, efficiency, and automation. How could these people, who existed on the cutting edge of technology, not figure out how to make that tech work better for themselves? I felt like, being so new to the industry, I must be missing something. So I did what I have always done when I need to clarify my thoughts: I started writing. I wrote about how you could automate the job of a VC analyst, the ideal tech stack for an emerging fund manager, and why I don’t believe in purpose-built software for VC. The writing helped me to clarify my perspective and also showed me that I was on to something. The ideas seemed to resonate with people. 

I also learned a tremendous amount through the process. I cultivated a beginner’s mindset and absorbed as much information as possible both through reading and through talking to people who had been in the space far longer than I. And, somewhere along the way, I became an expert. I could confidently say that I was one of the only people who was spending dozens of hours a week thinking about how to automate the operations of venture capital.

After six or so months of writing, I had developed a strong perspective, a strong base of knowledge, and a strong network of VC operators. When I joined Sydecar in 2021, I brought all of those things with me. There was plenty that I didn’t bring to the job: experience writing a mission and vision statement, developing a logo and visual brand identity, or managing an ad budget. I didn’t know what an ICP was, but I knew ours better than most people could. This meant that, in setting up Sydecar’s initial GTM strategy, I had to rely almost exclusively on my understanding of and conversations with people in our target market. I didn’t have any of the playbooks: SEO, ABM, PLG. I couldn’t rely on doing things the way I had done them before, because I had never done it before. There are a lot of choices I can attribute to that: I deprioritized a traditional SEO strategy in favor of editorial and opinion-based content. I went hard on Twitter and neglected LinkedIn. I rejected the idea of paid advertising for almost three years, but spent tens of thousands on sponsored events that I couldn’t prove ROI for. I did almost no product marketing and instead called customers constantly to answer their questions, get their feedback, or walk them through the product. I probably did a terrible job of justifying the decisions I was making because at the end of the day the answer would have been something along the lines of: it just feels right. 

So many of the fundamentals of B2B marketing have broken down over the past few years. Semi-personalized outbound, SEO, paid ads, LinkedIn executive content. Every channel is saturated, and every playbook has been run into the ground. Audience behavior has fundamentally changed (no more Google searching) and cannot be perfectly tracked and attributed like it once could. There is literally no shortcut to talking to and directly observing your customers. Which is why I believe that some of the best marketers to come won’t be marketers at all, but rather deep subject-matter experts. For one, these people likely have already built trust with your target market. They likely have deep networks within that market, which is invaluable as more intimate, network-based growth tactics, like customer referral programs and IRL events, become popular.

These networks also grease the wheels for customer discovery and feedback. As companies grow, marketers tend to become further and further removed from the customer. But if you don’t have the typical playbooks or tactics to rely on, then you’ll be forced to go to potential customers time and again when you have questions about how to prioritize budget or execute a campaign. 

A great example of this is economist Ara Kharazian of Ramp, who was hired last year to lead their economic data function. While not technically a marketing role, this has probably been one of the greatest growth unlocks for Ramp in the past few years. I wish I could think of more examples, but there aren’t that many (yet!). What I have observed is that tech companies seem to be more open to (or even in some cases to prefer) hiring from traditional industry jobs, especially when the tech they’re building serves that industry. Perhaps that’s an indication of further movement in this direction in the years to come.

It’s time for a change! We’re living in an age where just about every marketing and go-to-market tactic has been pronounced “dead”: growth hacking, SEO, paid social. What was once an obvious best practice, deployable by companies across verticals, now works for very few. Over the next year, I think we’ll see foundational principles of marketing break down alongside these individual tactics.  I see this shift happening most clearly with product marketing.

Summed up, conventional product marketing explains how a product functions in terms that resonate with a specific type of buyer, in hopes of convincing them that the specific offering is better than anything else out there. The emphasis is very much on how the product functions and who it is for. 

We’ve seen the pace of product development accelerate exponentially over the past year thanks to AI. Features that were once differentiated are becoming commodified. It’s no longer that impressive that your tool can automate a manual process in a fraction of the time. Claude Code can do that, and I’m already paying for it.

As this commodification of software continues, the job of a marketer stops being about positioning, messaging, or launching. It becomes making the case for a version of the world that needs your product to exist. This is why we’re seeing so many companies paying mid-six-figure salaries for narrative or storytelling roles. It’s not content marketing, it’s not brand marketing, and it’s not just PR or comms. It’s filling a gap that formed when we realized a product launch would never get the same number of likes, media mentions, or web visits as it did in 2021.

This will be accelerated as we move into a world where agents interact with software on our behalf. They will care about very different things than we do – for instance, clarity of copy and content structure. Pricing is a good example of this. Most companies have embraced the practice of including pricing details on their websites, but few do so with total transparency and clarity. As marketers, we feel a pressure to slap a shiny coat of paint on anything that we ship. So rather than displaying our pricing in the clearest possible manner, we spend hours workshopping copy to present it in the way that makes us look… the cheapest?

To be clear, a lot of this work requires the behind-the-scenes product marketing to be done, and to be done well. I continue to believe that a rigorous and detailed ICP definition is more important than ever before. Similarly, having a deep understanding of your competitive landscape and why you win and lose is more important than ever before. Listen closely to your customers. But, if you listen closely, you might find that a competitor is spending more time on opinion blogs and earned media than feature comparison pages, and that you’re winning more customers through referrals as opposed to your “Book a Demo” button. These organic, word-of-mouth channels do much better with narrative fuel than with traditional product marketing fuel. No one is talking about your most recent product release over cocktails. 

Like with anything, there are always counterexamples. I’ll continue to preach the importance of rooting GTM strategies in a deep understanding of your market and your ICP. If that research tells you to lean into conventional product marketing tactics, by all means do so. And let me know how it works?

In the decades since Supreme launched and made its name using this tactic, it’s been employed by everyone from denim brands to plushy toy makers to, of course, tech startups. And, because tech has to ruin everything, the tactic that once made Supreme a beacon of coolness to me has now lost its appeal. 

I generally believe that tech companies should do less navel-gazing and look outward for inspiration. I also believe that fashion brands are a great place to seek that inspiration because they tend to excel at visual storytelling, fast-paced trend adaptation, and balancing the aspirational with the relatable. FOMO marketing worked incredibly well for Supreme, and has worked well for countless clothing brands since – largely because 1) it allows startup businesses to gauge demand and manage their buying/inventory accordingly, and 2) when you are dealing in physical products, the idea that supply is limited is believable.  

But software companies don’t have inventory, nor are they selling physical products. So why have so many of them used waitlist marketing as a tactic? In some cases, it serves to gauge demand for a product not yet built. In others, it serves to build community or create lock-in amongst a group of early adopters. In far more cases, it’s a tactic lazily copied from a larger, more successful startup because someone doesn’t want to do the boring work of deeply understanding their ICP and building a motion that speaks to them. We’ll get into that in a minute – but first, I’ll give credit where credit is due to the handful of tech companies that have employed FOMO-based marketing thoughtfully and successfully. 

Robinhood 

Robinhood launched in 2013 with a simple landing page to collect sign-ups from people who were interested in commission-free trading. In this case, the waitlist makes sense because the product had not yet been built. It was an effective way for them to gauge demand for a risky concept in the aftermath of the 2008 financial crisis. At the time, products, especially complex financial products, took months to build. The waitlist made sense because users who expressed interest would have to wait to get access to the product being marketed. 

Robinhood.com; March 2014

Superhuman 

In the mid-2010s, Superhuman launched with its own version of a waitlist. In their case, they were gating access to a truly limited resource: 1:1 onboarding calls with their founding team. They used the onboarding as a way to ensure ideal fit with their early adopters, building on the idea that a friction-full onboarding experience can create a more bought-in user who is more likely to be a champion of your product and to make referrals (Superhuman’s early growth was almost exclusively referral-based). They kept up with the waitlist model until 2022, when they shifted to a self-serve, publicly available model.

Superhuman.com; February 2016.

Notion 

Then there was Notion in 2018. In their case, the list was a way for the company to establish a base of early adopters who were bought in from the beginning. Rather than treating their waitlist as just a marketing tool, Notion saw it as an opportunity to build their product in public, leveraging the list of beta users for product feedback and testing. The strategy makes a ton of sense for Notion, a company that has built its growth engine around the strength of community.

Notion.so; January 2019

Each of these companies used a waitlist to achieve a specific objective aligned with their go-to-market motion. For Robinhood, it was testing demand for a high-stakes, risky product that would take time to build. For Superhuman, it was creating a more bought-in user base. For Notion, it was establishing a legitimate sense of community before scaling.

In the years since these examples, hundreds (if not thousands) of tech startups have launched with their own version of waitlist marketing. Most of the time, it’s not entirely clear what purpose the waitlist serves. Last week, I saw a big AI company promoting the waitlist for a product launch that is happening… in the fall. THE FALL?! Don’t they know that the robots will already be in charge then?! 

Software doesn’t have a limited capacity. It’s easier than ever to build something — at the very least an MVP — to validate demand. And for the most part, companies aren’t taking the Superhuman approach, in which access is constrained by a limited resource like human labor. 

So, please. Do me a favor and take down the waitlist. 

And add your pricing to your website while you’re at it.

I was recently chatting with some B2B marketers, and the topic of marketing qualified leads came up. Specifically, how to define an MQL, how to ensure alignment between marketing and sales, and how to operationalize the scoring and outreach workflows. It’s a good time to rethink definitions and processes that once seemed obvious, especially as many of us rebuild with the help of AI tooling.  

While drafting my contribution to the conversation, I had this thought that marketing qualified might… not mean anything anymore? Or perhaps the issue is that it means something different to everyone, rendering the concept entirely useless. I don’t think the solution is to standardize it, because this misalignment is a symptom of something larger: the concept of marketing qualified was built for a world that no longer exists. A world where marketing’s job is to manufacture readiness. 

The original intent of the marketing-qualified definition was to identify prospects who were ready to be sold to. It was a solution for something that couldn’t easily be observed – it’s impossible to know that someone is fed up with their existing provider or that they got a new budget line item approved, both signs that they may be “ready” to buy your product. So, instead, we created proxies and assigned points to behaviors that correlated with readiness: downloading a report, attending a webinar, or opening an email.  

This entire framework assumes readiness is something marketing manufactures. (Side note: I think a lot about the term demand gen. Is demand something that marketing should need to generate? Shouldn’t demand exist for a product to be built in the first place?! Anyway, I digress.) 

This approach works (ish) for traditional SaaS, but it breaks down when applied to usage and transaction-based models. These businesses are becoming increasingly common in a world dominated by AI, but they existed well before and continue to exist well beyond AI. With these products – which I often refer to as “painkillers not vitamins” – buyers don’t move through a linear funnel just once, and they are often not in market at all. They enter and exit the market repeatedly and unpredictably, driven by their own circumstances. Readiness is an inconsistent and fluid state, not a one-time event. And it's mostly exogenous, created by the buyer’s unique set of circumstances, not your nurture sequence. 

Plenty of companies have begun to shift their GTM motions to better fit usage models. This most commonly shows up in pricing: tons of AI companies have adopted usage-based pricing rather than a subscription model, which allows them to cover costs, optimize revenue, and meet the buyer where they are when they need it. 

GTM teams are also eagerly adopting signal-based marketing, where you engage a prospective buyer based on an observation that they are ready to buy, such as the announcement of a new hire or round of funding. With the explosion of signal-based marketing, it’s important to remember that observed signals only matter if they’re rooted in a deep understanding of your ICP. Does the fact that a company hired a new salesperson actually impact whether they’d get value out of your software? Why? Beware of copying signal-based playbooks from other companies whose product and/or target customer looks nothing like yours. I understand the temptation to track every possible signal as first- and third-party data becomes more accessible than ever before, but this leads to lazy marketing. 

Marketers should learn to document key signals with as much precision and rigor as they do any other part of the ICP definition, like tech stack or media diet. Just as we used to care exactly which publications our ICP read and why, we should care to understand what signals growth or change within the organization.  Otherwise, we’ll end up with an explosion of data points and very little signal (pun intended) in the noise. 

Here are a few examples of what this looks like in practice: company name, signal mapped to ICP definition, and outreach tactics that leverage those signals:

Company

ICP

Signal

Why it matters

Tactic

Sydecar

VCs

Peer adoption

Buyers are skeptical and network-driven 

Referral program 

Clay

GTM strategists

New AE hire

Sales targets are increasing 

Enriched target list

Stripe

Startup founders

Funding round

They have new money

Boost their promo

Twilio

Product managers

Product launch

They need comms infrastructure

Free trial offer 

Shopify

Startup brands

Social media follower growth

Their sales are increasing 

Website builder demo

The funnel approach was more permissive of lazy marketing. You could skimp on knowing your buyer and defining your ICP because outreach was based on the idea that marketing could manufacture buying readiness. When you are selling a recurring subscription, the majority of buyers are always in market. As software companies continue to shift towards usage-based models, funnel-based marketing breaks down in favor of signal-based marketing, and signal-based marketing doesn’t allow us to be as lazy. In this new reality, the competitive advantage lies with marketers who are willing to do the specific, sometimes boring work of deeply understanding the people they sell to and the world those people inhabit.

Marketers have spent the past decade optimizing for frictionlessness. Whenever possible, we remove steps, pre-fill fields, and demonstrate value as quickly as possible. Any ounce of friction is assumed to be a place to lose potential customers.

This approach makes sense in a world where products compete solely on feature sets, and where a buyer moves linearly through stages of a funnel determined by a marketing team: "see an ad, read a case study, do a demo, buy a product." In that world, we want to move people as efficiently as possible from the top of the funnel (awareness) to the bottom (conversion).

But we are so far from that world today. Thanks to AI, it's becoming harder for companies to differentiate on features alone. We aren't choosing products based on features or functionality anymore. We're choosing based on how they make us feel, on whether they validate the version of ourselves we want to be, or want the world to see.

The friction worth keeping is cognitive: the work of learning a new vocabulary, building a new mental model, going down a rabbit hole on an unfamiliar concept. Not arbitrary gating, not FOMO-driven waitlists, not manual processes that exist because no one bothered to remove them. Buying behavior changes; human psychology doesn't. People still feel the pull of sunk cost after they've put real time into understanding something, and that effort is what turns a user into someone with actual affinity for the product.

That friction shows up in two forms. The first is literal product onboarding, or the experience of learning a new system. Take Claude as an example: to use it well, you have to set up connectors, learn how to prompt effectively, understand what an agent and a skill are, and know the difference between Chat, Cowork, and Code. Anthropic hasn't built an onboarding flow that holds your hand through setup, and I wonder if that's intentional. The learning becomes the product itself. Once you go through onboarding, you emerge with a shared language and a sense of belonging to a new group. 

The second form is conceptual onboarding. It's the rabbit hole someone goes down before they ever touch the product, most common in businesses like deep tech with long sales and education cycles. When the cycle is that long, and there's initially nothing to buy and no one to convert, conventional product marketing doesn't work. The work is educating the market, over years, on why your solution will be the right one when it finally exists.

Either way the mechanism is the same. Cognitive effort creates commitment, and commitment is the difference between a user and a champion.

That champion matters more than ever, because buyers no longer take the company's word for anything. They want to hear what a product is worth from someone they already trust. Social proof used to come from the company: logo walls, testimonials, badges, case studies. That kind has been collapsing for years, and AI has finished the job, since everyone now assumes anything on your owned channels was ghostwritten or fabricated. What's left is earned social proof: a Reddit post where someone shows what they built, a YouTube tutorial that teaches the product better than the company can, a Discord conversation between users who don't know the company is watching. The customer who fought through friction is the same one who is producing social proof in the public sphere to signal something about themself.  

What bothers me about this line of thinking is that it treats AI as a productivity layer that is being applied to a static landscape. In reality, AI is changing the landscape so fundamentally that old functions shouldn’t be automated, but done away with. The same technology that makes a voice agent possible is also changing how buyers research, evaluate, and decide. They're doing more of the work themselves, earlier, and asynchronously. They expect the people reaching out to them to already know who they are and what they care about. Cold calling was only ever tolerated as part of the cost of doing business in a world with limited information signals, but we don't live in that world anymore.

Rather than using agents to make more calls (or send more emails, or publish more ads), the actual opportunity sitting in front of GTM teams lies in the explosion of signals. It’s easier than ever before to understand who's reading what, who's in market, who's been to your site three times this week, or who's talking about your brand in their Slack community. The value isn't in automating the input, but in learning to use the signals to be more precise about when you show up and what you say.

There's also a shift happening around permission. Outbound used to operate on an implicit permission: you were a prospect, so outreach was expected. That's no longer true. Permission now has to be earned by being useful before you ask for someone's time, by showing up in the places they already pay attention, or by being championed by people they already trust. There’s no silver bullet or shortcut for any of these processes, via a voice agent or any other sort of AI.

The moat doesn’t lie in optimizing or automating old motions, but in building new ones off the back of the opportunity that AI creates. Good marketers today are embedding themselves in the communities their buyers are already a part of and partnering with organizations and people who've already earned trust in those communities. They're producing things that their ICP actually wants to consume. AI certainly plays a role in developing these new motions, but it’s not a replacement for deeply understanding your customers. 

Doing this right requires patience and conviction, which is the hardest part for most organizations. The ROI of owning your own community, or executing a large-scale brand activation, is lagging and hard to capture in traditional attribution models. The work of identifying which organizations have already built trust with your ICP, and figuring out how to partner with them, requires sustained attention and a fair amount of taste. It’s certainly much harder than buying a lead list or running a paid LinkedIn campaign. But, at the end of the day, it’s marketing’s job to tell the story about why this “slower” work is so important. Teams that make the case effectively will build real moats as their work compounds over time. Once you've embedded in a community and built relationships with the people who matter in it, you have something competitors can't replicate.

Inspired by Melissa, this week’s newsletter is going to be my own “Good List,” focused on some percolating thoughts that I’m not convinced warrant a full 800 words to themselves. 

Domino's Pizza Tracker

Someone recently suggested adding a “Domino’s Pizza Tracker”-type feature to our product to let customers track the various stages and statuses of their SPVs. I’ve always been a fan of this feature, despite learning the hard way that it’s fake while trying to order a pizza at 2 AM after a night of drinking in college. The conversation spurred me to dive into the tracker’s history: it was introduced in 2009, when Domino’s, following a devastating taste test in which they tied for last place with Chuck-ee Cheese, hired a new CEO to revamp its brand. The new brand was built around transparency as a way to communicate to the public that they were investing heavily in getting better. They executed a campaign that included ads, a full-length documentary, and – you guessed it – the introduction of the pizza tracker on their website and in their app. It worked, and their sales and stock prices soared in the subsequent months and years. This is a great lesson in the value of simply doing what you say you’re going to do and bringing customers along for the ride. Whether or not Greg is actually the person assembling my pizza, I still get peace of mind from the notion that there is someone for me to hold accountable if it doesn’t get delivered. 

Jacquemus Valerie Bag campaign

I’m so tired of seeing the same tired takes on the state of B2B marketing on LinkedIn, so I’ve started to look to other sources for inspiration. I’ve been spending a lot of time on Pinterest (a phenomenon that probably warrants a full article of its own), specifically looking at fashion brands.

Earlier this year, Jacquemus, a luxury women’s brand from France, released a campaign for one of its new bags, Le Valérie. Jacquemus has historically taken the tonal, restrained approach that we’ve come to expect from a luxury brand, but this campaign felt quite different. It was deeply human and even a little sloppy. It was a clear attempt to relate to the everyday busy woman without losing the clean, premium, aspirational feel that the brand is built on. They made a clear decision about who they wanted to speak to and how they wanted that person reflected back at her, and every choice in the campaign served that decision. This isn’t the first brand to intentionally introduce a bit of human slop into its marketing as a way to counter the rampant AI slop. I’m excited to see more versions of this in the near future.

The campaign is also an acknowledgment that people buy based on the beliefs they hold (or want to hold) about themselves, not just based on the product. This has always been true in fashion, and it's increasingly true elsewhere, including software. The barriers to building products keep dropping, and products keep converging. If someone can vibe-code their own CRM in a few weeks, the question of why a customer chooses Salesforce or HubSpot won't come down to features. It's going to come down to the story they want to tell about themselves. The Jacquemus campaign was a clean example of a brand understanding that and building the entire creative around it.

Spotify 20th Anniversary Campaign 

Spotify as a company confounds me. Of course, there is a lot to admire. They are one of the greatest case studies of building viral loops with their iconic Spotify Wrapped, which is top of mind this week since they launched the “Spotify 20: Your Party of the Year” campaign in honor of their 20th Anniversary. Similar to Wrapped, the campaign encourages users to share artifacts like their first-ever song played on Spotify or the exact date they joined the platform via Instagram or other social media accounts. Just south of 90% of Spotify’s revenue comes from subscriptions, so campaigns like this that leverage other, larger platforms for user growth make a lot of sense.

But there is still so much left to be desired from the platform, and I have to imagine that is hurting their retention numbers. Firstly, I don’t understand why they have never leaned into the social aspect of the app. I, like many people I know, love to learn about new music from my friends (of the real and internet variety). It’s why I used to love Soundcloud, and why I still spend a good amount of time on the platform. Soundcloud is far from perfect, but each time I log in, I’m greeted with a feed of sounds curated by people I’ve chosen to follow. Spotify, on the other hand, has no feed feature to speak of (unheard of in 2026!). 

Spotify’s recommendation algorithm is also atrocious. Algorithms should be used for discovery! I don’t want a playlist of songs or artists that I listen to regularly and that are already on playlists I have created for myself. Given the amount of data that Spotify has on my preferences, it seems ludicrous that they aren’t able to help me discover new music that I enjoy – and that they aren’t even aiding my discovery process by encouraging me to follow people who can curate new sounds for me. The disco ball campaign is cute, but it’s not enough to keep me from spending more and more time on Soundcloud or NTS.

Of course, AI makes this more difficult. The digital landscape, where most marketers have spent the majority of their time over the past two decades, is changing faster than ever before. And, as it turns out, AI is reasonably good at a lot of the tasks that we used to be responsible for: writing copy, identifying target buyers, generating images. Reasonably good is fine for most of marketing, the same way it's fine for ordering a beer at a busy bar when you really wanted a cocktail. The challenge for marketers today is identifying where 90% isn't good enough, and where the last 10% is something people will actually pay for, either with money or with attention.

I’ve been thinking a lot about curation as the next frontier of opportunity. On the surface, curation is a killer use case for AI: it's just an algorithm applied to a large dataset, which is one of the things AI does best. Curation has been a core feature of the internet for decades: it’s what made Tumblr good, and what makes Twitter bad now. Algorithmic curation is also what lets us spend less on audiences and more effectively reach the right audiences. Ad platforms have gotten dramatically better at this as they've used AI to parse through first and third-party data, intent signals, and behavioral patterns. 

But interestingly, the more algorithmic curation becomes, the less I find myself relying on it. I can’t stand my Twitter feed, but I am subscribed to more newsletters than ever before and often refer to the writers as though they are my friends. I genuinely couldn’t fathom taking a restaurant recommendation from Claude, especially when my Google Maps is teeming with pins that I’ve painstakingly collected from friends and strangers alike. I love Pinterest, but I have made a purchase from a sponsored post.  Since moving to DC, I’ve delighted in the experience of asking new friends for recommendations for primary care, dental, therapy, dermatologist, haircut, botox, etc. Zocdoc could never. 

In all of these cases, it’s not a matter of the curation being more “accurate,” it's that the recommendations come with taste, point of view, and something at stake. While AI curation optimizes for engagement, or relevance, or some proxy, human curation optimizes for "I would feel weird if I sent you somewhere bad." Those are different functions, and the second one is hard to replicate because it requires the curator to be accountable to me in a way an algorithm structurally can't be. 

A better recommendation engine doesn't get you closer to your friend's restaurant rec, because the value of your friend's rec isn't that it's more accurate. It's that she ate there, she knows you, and she'd feel weird sending you somewhere bad. 

As AI collapses the cost of generic curation to zero, the only curation worth paying for (or paying attention to) is curation that carries someone's taste and reputation. We can see this playing out clearly with events as a marketing channel. Following the post-COVID event boom, people (especially those in SF, NYC, LA, Austin, etc.) became much more selective about what events they would attend. The annual tech weeks went from a hot ticket to a necessary evil to something to be avoided altogether, simply because both the events and attendees lack thoughtful curation.  In contrast, intimate event experiences with curated guest lists are on the rise. 

The thing that makes these events work is not the venue or the food or the talking points. It's that someone has decided who's in the room, and that decision is legible to everyone there. It’s the idea that each attendee is being hand-selected by the host because their attendance will make the event better for everyone else. You're at the table because the host wanted you at the table. Personally, I’d love to receive more event invites where the host calls out 2-3 attendees I’d get value from connecting with.

So who’s throwing the next party?

Queue, “And here’s what that taught me about B2B marketing.” (Sorry.)

AI has done to written content what GLP-1s did to weight loss. A junior marketer with a Claude subscription and a dream can pump out dozens of articles a day with the perfect format, structure, and keywords to hit their target audience. Which means, predictably, that content-driven marketing tactics (SEO, the company blog) have started to lose their power. There are still companies that are seeing growth through traditional content channels, but this is becoming the exception simply because it's far too easy to create marketing content. No one (hyperbole) is making meaningful buying decisions based on a five-word Google search. They are looking to recommendations from an AI that has context on their preferences over generic search results, and to peer recommendations over a testimonial on your home page. Emailing a peer for a recommendation may take longer than reading a testimonial, but it results in a recommendation that we feel much more confident in. We understand, on some level, that nothing worth having comes easy.

This is also why we’re seeing more companies standing up media arms, launching podcasts, and doing video. Of course, part of it is that AI cannot (yet) create high-quality video. But there’s also an understanding from the audience (conscious or not) that these forms of media take a lot of work, and we favor companies that will put in that work. Plus, if a company has the resources to do video, they're probably well-capitalized, and a buyer would rather work with a company they know has longevity.

Tanay Kothari, the CEO of Wispr, recently shared that for their India launch, they wrapped 100 rickshaws in Wispr branding and sent them into Bangalore traffic. He points out that American tech companies are burning through millions on digital ads to reach Bangalore's tech community, all bidding on the same keywords and running identical LinkedIn campaigns, while the physical attention of an entire city sits in gridlock, staring at the back of a rickshaw.

For most of the 20th century, if you wanted to reach a mass audience, you bought physical space (billboards, print, transit ads). As demand for OOH advertising increased, it became more expensive (and also difficult to measure in terms of ROI and attribution). When digital advertising came along in the early 2000s and scaled through mobile in the early 2010s, it offered a lower entry point along with better targeting and attribution. Advertisers eagerly flocked to Google Adwords and the Facebook ad network because they were the cheaper alternative.

In the past decade, digital advertising has followed the same trajectory as the billboards it replaced: advertisers flocked, and costs climbed. As the internet has become more complex, targeting and attribution frameworks have also broken down. All of a sudden (20 years later), the cheaper alternative has become the more expensive one. In recent conversations with other marketers, OOH, direct mail, and in-person events come up more often (and with more excitement) than any digital ad channel, lead magnet, or AEO tool.

It feels like the channel that is most effective at any given time is the one that requires the most effort, or the most creativity, or the most resources. Not because effort is inherently virtuous, but because it is self-selecting: most people won't do it, which means the ones who do will have less competition.

I spent a lot of time writing SOPs (Standard Operating Processes, for the uninitiated), covering things like how to log a discovery call in Salesforce, how to move a deal through stages, what counts as a qualified lead. I organized them carefully and kept them current, always excited for the opportunity to put a new “Last Updated” date at the top of a page.

I had a lot of pride wrapped up in this work, maybe more than was warranted. Here I was, this hardworking, organized, ambitious young person doing a thankless job in a world where slacker salespeople were always going rogue. Looking back, I think I was playing the martyr a little. The documentation existed partly because it needed to exist, and partly because creating it made me feel like I was doing something real in a job where my impact was hard to see.

Regardless of the impact I was (or wasn’t) having at the time, I’m grateful for the opportunity to build good habits around process design and documentation. These are skills I’ve carried into my career as a marketer. The first person I hired at Sydecar wasn't a marketer or a salesperson. It was a generalist with a data and ops background, because I wanted to build the tracking infrastructure before we needed it rather than after we wished we'd had it. This is not a particularly glamorous hiring decision, but it’s one that I stand by.

Every year, in the quiet stretch between Christmas and New Year's, I write a long document about the year behind and the year ahead. Most of the company is out and nothing urgent is happening, so I use this time to pull data, revisit decisions, and get honest with myself about what worked and what didn’t in the previous 12 months. The resulting document is usually long, so much so that I typically know while I’m writing it that it will never be read in full by anyone else. When I eventually share the strategy with my team, I turn it into a deck with a few bullets per slide, because I know a conversation will land better than a wall of text.

I used to think that the documents I created were just for prep, and the real deliverable was the deck, the presentation to leadership, or even the actual execution of a strategy over the subsequent months. I’m not sure that’s right anymore.

AI has changed so much about how we operate as marketers. I will spare you from the exclamations of “the company blog is dead” or “Did you know that Reddit is the number one cited source by LLMs?!” What I’m more interested in, as I think about how AI has changed the role of a marketer, is the increasing importance of context. It’s truly the most valuable thing you can give a model. When I say context, I’m referring to the accumulated logic of how you think, what you've tried, what constraints you're operating under, what words you do and don't use, what your north star is and why. In the absence of this context, AI will act like a 2010s management consultant and spit out generic playbooks or overplayed tactics (I need an entire article to talk about my hatred for playbooks).

As a remote employee, so much of who I am as a marketer is formed somewhere between my brain and the keys of my computer. That means I am constantly documenting context (across a wide variety of sources, including email, Slack, and Zoom calls) on business decisions, strategy, direction, projects, learnings, etc. Creating context documentation is both easier and more important than ever.

This feels like a meaningful inversion of how we've thought about documentation. For most of my career, the context doc was the thing you did before the real work. Now I'm not sure the real work is ever the output. If the model is producing the first draft, whoever shapes the context it works from is doing something more consequential than editing copy.

I've started to notice that the people best positioned for this shift aren't necessarily the career marketers; rather, it’s the people who are good at articulating context. People who can write down not just what they want but why, what's been tried, what the constraints are, why it matters. I don't know if this is a hot take, but I'd bet on the ops person before I'd bet on the creative. They’re the ones who were maintaining documentation long before anyone was going to read it.

And the good news is, models don't get tired of reading. They don't skim, and they don't show up to your strategy document having already formed an opinion. They consume everything you give them and they apply it, which means the quality of what you give them actually matters.

I was there for a conference put on by The Idea Village, the same organization where I had my first marketing internship in 2015. And it was hosted at Common House, a members club in the Warehouse District that used to be the Louisiana Children's Museum, which happens to have been my first post-grab. I stayed at the hotel on site, and I'm almost certain I was sleeping in the exact location where my office used to be.

I generally liked that job. It only required my attention between 9 and 5, which left a lot of room for being 22 in New Orleans. But there was a lot I didn't love about it either. I was the only full-time marketer on staff. My boss was a fractional comms/PR woman, who cared enormously about brand and not at all about data, demand gen, or trying literally anything new. I remember asking her for a small experimental budget to test Facebook and Instagram ads (we were marketing to millennial and Gen X parents, it seemed so obvious). She told me there was no advertising budget available because it had all been allocated to ads on the historic New Orleans streetcars.

I tell that story a lot, mostly to illustrate why I made the transition from non-profits to tech startups (or went to the dark side, as my parents would say). It’s funny to think about now, in a time when Meta ads are so passe and OOH is once again in vogue. Was just simply ahead of her time?

I don't think so, exactly. But I also don't think I was as obviously right as I felt. I was 22 and freshly convinced that digital was the only thing that mattered. Social ads were underpriced then in a way they aren't anymore, so I wasn't wrong. But there's a difference between being right and understanding why you're right, and I definitely didn't have the second one.

I've been thinking a lot about how much has shifted in the ten years since. Not in a "the landscape has evolved" way, but rather that a lot of what felt like settled wisdom then has been walked back, and a lot of what felt dated has come back around. I don't think the takeaway is that everything is cyclical or nothing matters. I guess I'm still working out what the takeaway is, so I’ll have to write another one of these at some point. Will try not to wait 3 years this time.

For my first ten months at Sydecar, the company was in stealth and the team was still very small. I was the first and only “business” hire and many of the lessons that I learned were learned on my own. I didn’t do a good job of finding mentors or building a support network around myself. I am confident that because of this, I made more mistakes, spent more time questioning my judgment, and took longer to make important decisions. 

In contrast, I saw people around me in the startup community asking questions, admitting mistakes, and showing general vulnerability in very public ways. Sometimes, this made me cringe – and sometimes, it made me deeply jealous. Frankly, the notion of “building in public” has never resonated with me. Maybe learning these lessons the hard way was actually important for me. It required grit, resilience, and fighting off a tremendous amount of imposter syndrome. In some ways, I guess I am better for it.

But I appreciate the intention of building in public. No doubt, it is an effective way to learn quickly and build a robust support network. While I am by no means pivoting to a “build in public” person, I am committing to sharing more of my thoughts and learnings publicly in 2023. If there is a single person who can learn from the mistakes that I made, then I think it’s worth it to share them. Also, I’ve done very little writing for myself over the past 18 months. I know that writing is a phenomenal way to reflect on both mistakes and successes and continue learning from them. I hope it’s also helpful, or at least entertaining.