How to Grow in the Age of AI

Part 1: Exploring Acquisition Loops: Viral, Content, Paid and Sales

Hi there, it’s Adam. I started this newsletter to provide a no-bullshit, guided approach to solving some of the hardest problems for people and companies. That includes Growth, Product, company building and parenting while working. Subscribe and never miss an issue. If you’re a parent or parenting-curious I’ve got a podcast on fatherhood and startups - check out Startup Dad. Questions you’d like to see me answer? Ask them here.

---

Do you know a Startup Dad or Mom who could use a little more merch in their life? Check out: www.startupdadshop.com.

Thanks to Mister Branzino for his purchase!

---

Q: How does AI change my growth strategy?

Regular readers of this newsletter will know that one of the ways I spend my time is teaching the Growth Series, Advanced Growth Strategy, and Growth Leadership (which I co-created with Elena Verna) at Reforge. Also, unless you’ve been living under a rock you’ll know that AI is the new hotness. I’ve generally stayed away from writing about it lest I be associated with some sort of prompt-engineering broetry on LinkedIn. I am, however, an avid consumer of a handful of standalone AI tools and some of the newly released features within my favorite products. I use AI in Notion and Canva and have subscriptions to ChatGPT, Perplexity and Midjourney. I’ve also explored a lot of AI tools as an investor. 

Lately, I’ve been fielding a lot of questions about the role of AI in growth strategy and also thinking through how certain growth loops will be enhanced or decimated by our AI overlords.

So I decided to write about it. Over a few newsletters I’ll break down the core components of growth—acquisition, retention and monetization—and share some thoughts on what I think might happen with each of them in the age of AI!

• Acquisition in the age of AI ← this post

• Retention and Engagement

• Monetization

Acquisition in the Age of AI

There are four main types of acquisition loops - a viral loop, a content loop, a paid loop, and a sales loop. I’ve written about different examples of these loops in the past, like templates, profiles, and invites.

Recall that an acquisition loop involves an input, an action, and an output that can be reinvested as an input. That’s what makes it a closed loop system.

Let’s explore the different types of loops and how they might change with AI.

Viral Loop

A viral loop involves someone telling or inviting someone else to a platform. If I bump into you on the street and say, “Hey, you should check out Adam’s FishmanAF Newsletter!” that’s a word of mouth viral loop. There’s no trackable link; it’s just me telling you about it.

If I use Substack’s native functionality to generate an image and a shareable link to the FishmanAF Newsletter, that’s an Organic Viral Loop. I shared a link with you out of the kindness of my heart and because I wanted you to subscribe.

If I use Substack’s referral flow to share the newsletter with you in the hope that I will receive benefits and rewards, that’s an incentivized viral loop. 

And if I started some sort of live Zoom chat to talk about Growth and in the corner I had the FishmanAF Newsletter logo, that would be a “Casual Contact” viral loop.  

All good with our terminology? Great. Let’s move forward.

How will AI help or hurt viral loops?

I think it’s a mixed bag in terms of impact on Viral Loops. For the Word of Mouth and Casual Contact sub-types, I don’t think anything changes. Unless someone generates a deep fake of Taylor Swift telling her followers how amazing the FishmanAF Newsletter is; I won’t see a whole lot of benefit. Same with that logo hanging out in the background. I suppose we could generate more interesting backgrounds, but doubtful that changes much in terms of viral loop performance.

For invites and incentivized invites I do think some things could change. The level of sophistication required to generate an individually personalized invitation will go to zero. Feed in a small amount of information about the recipient and I can create a very customized recommendation and message for you. Opens, clicks, and ideally signup conversion (from a highly personalized and individual landing experience) should all improve.

But Adam, isn’t this possible without AI? 

Kind of, and there are several companies who create personalized experiences today by including the face of the person who invited you, or using your first name in the outreach. But I expect two things will happen: first, it’ll get a lot easier to do this and second, the level of personalization should increase dramatically. If I were one of the off-the-shelf referral and invite platforms I'd be scrambling to include this functionality so I don’t go bankrupt.

The other aspect of the viral loop that could change is the branching factor. A branching factor is the number of people a platform is shared with each time it’s shared. This number could increase with the help of machine learning (which is a subset of AI) through better recommendations. If machines are somehow able to better select the population you should be inviting, they might capture some people you didn’t think of and increase the branching factor for invites. This isn’t necessarily new though as companies like YesGraph (acquired by Lyft) were doing this as far back as 2012.

I’d give this a 5-out-of-10 in terms of impact to viral loops and their subtypes.

Content Loop

Spoiler: for the Content Loop I think it’s going to be a lot worse.

A content loop involves creating and distributing content, not the platform itself. Content can be generated or distributed by users or the company itself. That is how we define the subtypes. For example, the template loop is initially a form of company generated and company distributed content. Eventually, this expands to user generated content as well in cases like Canva, Notion, Miro, and more. 

Something like the guide library that Intercom publishes is a form of company generated, company distributed content and images and videos I share on Instagram are user generated and user distributed content.

The primary distribution mechanisms for company distributed content are search engines and email. The primary mechanism for user distributed content is social media.

Now that we’re good with terminology and distribution mechanisms, let’s explore the coming content apocalypse via AI.

How will AI help or hurt content loops?

As I mentioned in my spoiler above, I think that many content loops are going to be absolutely decimated by AI.

The rise of LLMs will collapse the cost and time associated with a lot of company generated content to zero. We’ve already seen AI integrated into nearly every single writing tool on the planet. You also now have standalone tools like Jasper.ai or API-driven AI content creation via tools like Content at Scale. All of these—plus the proliferation of image-creation tools like Midjourney, Canva, and DALL-E—promise to make it infinitely easier to create more content.

Adam, doesn’t this mean it’ll be easier for people to build content loops?

Not quite. It’ll certainly be easier for companies to create content, but that won’t necessarily make loop creation easier. And that’s because of the second impact of AI: search engine results pages.

Starting back with Featured Snippets, Google started to introduce what people refer to as “zero click search results.” This means that the answer to your search query is contained on the page. 

Like so:

This is the beginning of a slippery slope as search engines continue to experiment with generative AI search results and show more answers vs. links on the page. The one thing keeping this at bay for now is the monetization model. Google still doesn’t know what to do with all those sponsored links when it’s giving you answers on the SERP.

But, it may not matter what Google thinks (I never thought I’d say that). The reason it may not matter is because Google could get disintermediated entirely. I’ve heard from more early adopters who are turning off Google in favor of tools like Perplexity, ChatGPT, and others. And just last week the Browser Company, makers of the Arc Browser, demoed a new version of Arc that skips Google altogether. It’s still early, but you can see where this is going.

On the one hand content generation will get easier and easier, but on the other hand the wayfinding mechanisms of the internet might make that significantly less useful for top-of-funnel and information-providing written content.

However, there is some content that will remain safe. Content like templates and profiles—the type of content that requires a visit in order to complete a transaction—should be fine for now. In fact, that may even get better with the enhancements of AI (more, specialized templates and personalized profiles). User generated content that feeds the social media machine should also be fine for a while and continue to send sites traffic (albeit less and less because of organic reach, but that’s a different problem). Unfortunately, all those company generated Intercom resources I cited above? Soon to be swallowed by the AI monster. 

There is one interesting approach that companies could take and that is to create their own GPTs trained on their library of content. This could have the positive benefit of providing links back to specific source material, but we’ll still have the problem of finding the “Intercom Bot” (for example). So while this is interesting I don’t expect it’ll be something that growth practitioners can exploit.

Clearly though, a reckoning is coming for content. I’d give this an 8-or-9-out-of-10 in terms of AIs future impact to content loops and their subtypes. 

Paid Loop

This loop needs very little introduction. If you’ve purchased ads online you’ll know what a paid ad loop is—provided that your returns are quick enough to be reinvested into more ads. The online advertising industry is approaching a trillion (!) dollars in the next few years. 

And now it’s being completely reinvented with AI!

How will AI help or hurt paid loops?

For years the limitations for advertisers on Google, Facebook, Instagram and more have been their ability to create and optimize the performance of ads. The ad platforms know this and have been steadily incorporating machine learning for targeting. Recently they’re moved on to full-blown creative production powered by AI. Everything from Responsive Search Ads and AI Video Ads in Google/Youtube to Meta’s AI Sandbox and their rollout of generative AI features for all advertisers. Pair that with the dozens (hundreds?) of companies helping you build landing pages via AI and we have a paid loop revolution on our hands!

I’m optimistic that AI makes ads better, but I haven’t seen it yet in my own feeds and search results. I do see a lot of ads for AI tools on Twitter (X?) so that’s something I suppose.

Much like Content Loops there is one significant way that AI can impact paid ads for Search and that is what happens to the SERP itself. Google has been loath to clip the wings of its golden goose (sponsored listings) but there could be a world where the real estate for paid ads decreases. This could then have a negative impact on pricing for advertisers and reduce the efficacy of your paid search ads.

On the other hand, provided it doesn’t royally mess up or hallucinate, an LLM can generate a near infinite supply of ad copy and highly-personalized, individual variants in a way that humans cannot. Responsive search and display ads, Google’s ad copy score and dynamic search ads have been hinting at this for a while now.

The same is true for image- and video-based ads. Meta is just dipping its toe in the water with backgrounds, headlines and image expansion. Same with Youtube and their AI-powered ad creator tools. We’re not too far away from fully generated ads based on a quick scan of your site, app, or an uploaded product catalog.

There are two main risks I see with this. First, the audience and impressions available for ads are not growing as fast as the ability to create them. This has always been a problem and will just be further exacerbated by unlimited ad generation from AI. Second is the commoditization of ad creative. It’s bad enough that every direct-to-consumer clothing brand looks identical to each other in advertising, but what happens when everything else starts to look and sound the same? It’s safe to say that it probably won’t be good for your paid loops.

I thought that maybe we’d see more unique influencer-type campaigns but given that we now have AI-created influencers I’m not sure that this is true anymore.

The influence of AI on paid loops seems like a 7-or-8 out of 10 in terms of impact to me.

Sales Loop

Our last loop type is the sales loop and if I were a BDR or AE I’d be a little nervous at the moment.

Sales loops are where a human (sales rep) works a pipeline of leads to convert them into new customers. The profits from that are then reinvested into more sales reps forming a loop.

How will AI help or hurt sales loops?

We’re already seeing the impact of AI on sales loops. Standalone tools like Clay or new AI functionality added to existing tools like Hubspot, Salesforce or Outreach promise to increase rep productivity and automate away a lot of repetitive tasks.

I haven’t been shy about my thoughts on sales outreach. I’m hopeful that integrating AI into sales workflows will improve the relevancy and targeting of communications I receive, reduce the volume of contacts, and make the content better. But I actually think it’s just going to get a lot worse for a while. With current ML models we’ve had the tools to make email targeting and frequency better for a long time and yet very few companies seem to avail themselves of this magical technology. It’s quite possible that the robots will decide I need 13 sales touches instead of 7 and find me in more places. This makes me nervous as someone on the receiving end. 

From the company perspective, I do think that AI will make the performance of sales loops better. Used correctly, the content and targeting should both improve. This should save sales reps a lot of time and companies a lot of money, but also runs the risk of eliminating a lot of lower-level sales jobs.

Definitely an 8-or-9-out-of-10 on impact on this one.

Wrapping Up

We can’t avoid it, the AI robots are coming for our growth models and motions. I believe that the impact will be fairly large, not just on acquisition but also on retention (next topic!) and monetization.

Across the four loop types discussed today—viral, content, paid and sales—we should see AI have the least impact on the viral loop. Machine learning has already existed to optimize invite selection and an increased level of personalization probably won’t have much impact on invite performance. Barring some sort of deep fake I don’t believe word of mouth will change much for companies.

The other three loop types will see profound impacts. 

Content creation cost will trend to zero and top-of-funnel company generated content will get absorbed into search results pages or disintermediated by GPTs or browsers themselves. Existing content types like templates and profiles should continue to matter since they are on the critical path to a transaction. They can actually be made better by AI’s ability to personalize site experiences.

Paid loops have already begun their steady march towards incorporating AI improvements. Machine learning began powering ad targeting several years ago and recent advances from Google and Meta are bringing generative AI to ad creation. The robots haven’t taken over entirely yet, but we’re headed there! It remains to be seen what the housing for advertising looks like, especially on SERPs, and also what the impact is when an infinite supply of advertisements meets a non-infinite supply of eyeballs.

Finally, like paid ads, sales loops should benefit from better content, targeting, and response speed. Costs will come down and fewer people will be needed to do the work of BDRs and entry-level AEs. Good for companies, not-so-good for employees I imagine.

In my next post on the Age of AI I’ll explore habit formation and retention, dive into what onboarding looks like with the help of AI, and deconstruct some of the current chat experiences including how to build user habits with the most horizontal products we’ve ever seen.

Until next time!

_

Valentine’s Day is right around the corner and what do you get for the Dads and Moms who have everything? Startup Dad merch!

Check out: www.startupdadshop.com and be the hero this year.

-