TL;DR: As generative AI matures, HubSpot’s “all-in-one” software for marketing, sales, and customer support has an opportunity to evolve into AI-powered flywheel that accelerates (and eventually automates) every aspect of a business’s relationship with its prospects and customers.
- Historically, HubSpot has focused on the “middle of the funnel:” attracting and converting potential customers with educational content and automated email campaigns.
- HubSpot has evolved into a complete CRM suite, with apps for every stage of the customer relationship.
- With its acquisition of Clearbit in 2023, HubSpot is well positioned to build a AI-powered flywheel that simplifies the process of attracting, converting, and supporting the “best-fit” highest-lifetime value customers for any B2B product.
HubSpot’s Big Idea
When Brian Halligan and Dharmesh Shah founded HubSpot in 2006, Google Search was solidifying its role as the nexus of the internet, YouTube was still in its infancy, Facebook had just started expanding beyond college students, and Twitter was not yet a thing.
In this context, HubSpot’s Big Idea was Inbound Marketing. The narrative went like so:
Before, marketing and selling were dominated by outbound tactics—TV and radio ads, direct mail, and cold calls—that interrupted people’s lives.
In the new era, these intrusive tactics were becoming less effective. Instead, businesses needed to make it easy for potential customers to find them when they were:
- Looking for solutions to their problems
- Seeking relevant knowledge
- Exploring new opportunities
To achieve this, you need a new strategy: Inbound Marketing.
Inbound Marketing focuses on creating valuable, educational content instead of traditional digital brochures:
- A blog filled with shareable, SEO-optimized posts
- A library of educational videos, webinars, and PDFs offered in exchange for contact information
- An automated email marketing system to follow up with prospects
By executing Inbound Marketing effectively, businesses could ensure their content, brand, and products appeared in relevant search results and their prospects’ newsfeeds.
The goal: earn attention, convert it into interest, build trust, and ultimately make the sale.
HubSpot provided the all-in-one software to create and manage this comprehensive inbound marketing approach and supplemented it with a growing library of educational resources on every aspect of it.
HubSpot’s Big Idea was a visionary response to the internet’s transformative impact on marketing and sales. And it set the stage for its evolution from marketing software into a solution managing every aspect of a company’s relationship with its customers.
The evolution from all-in-one marketing to all-in-one CRM
HubSpot’s Big Idea was not just an insightful observation about the internet’s transformative impact on marketing and sales; it was the foundation of a new approach to business growth.
But for the first six years, HubSpot’s software lagged behind its ambitious vision.
Initially, HubSpot’s tools worked well enough for small businesses that lacked the technical know-how or resources to build websites, host blogs, and integrate lead-generation forms, CRMs, and marketing automation apps.
But the software was often clunky and lacked the sophistication required by more complex organizations.
From 2009 to 2011, HubSpot attempted to move upmarket, targeting small enterprises and B2B tech startups.
These customers needed reliable two-way sync with Salesforce and seamless integration with other apps, but HubSpot’s software at the time fell short.
It was plagued with bugs and lacked the robust features necessary for larger, more sophisticated operations.
The clarity of its founding vision and the excellence of its marketing collided with the limitations of its software.
Around this time, David Cancel and Elias Torres, serial entrepreneurs with their own marketing software company, Performable, had their own insights about B2B marketing.
They recognized the complexity introduced by new distribution channels, varied tactics for converting traffic into leads, and the multiple personas involved in B2B sales decisions.
Performable aimed to answer essential questions for B2B marketers:
- How to identify which marketing content contributed to successful sales
- How to get credit for those results
- How to turn static websites, landing pages, and email campaigns into personalized experiences tailored to individual prospects
Performable’s vision was ahead of its time: the range of supportive technologies wasn’t yet mature in 2011.
HubSpot acqui-hired Performable for $20M, bringing Cancel and Torres on board as HubSpot’s Product Lead and VP of Engineering, respectively. They spearheaded a major overhaul of HubSpot’s product and engineering functions, rebuilding the foundation of its codebase.
Over the next 12 years, HubSpot expanded from its renewed foundation into a comprehensive customer relationship management platform.
It evolved beyond marketing to encompass sales, marketing operations, and customer support.
Now, in May 2024, HubSpot offers a suite of individual apps:
- Marketing automation (Marketing Hub)
- Content management and publishing for websites, landing pages, and blogs (Content Hub)
- Sales pipelines (Sales Hub)
- Marketing and sales operations and data integrations (Operations Hub)
- Customer support (Support Hub)
- B2B payments (Commerce Hub)
HubSpot’s $150M acquisition of Clearbit in 2023 added a leading B2B customer data enrichment product to its portfolio. Meanwhile, the company has been steadily integrating generative AI features into all its core apps.
This sets the stage for HubSpot to become an AI flywheel that automates every aspect of the customer relationship.
The emerging risk to Google’s business model
In early April, word spread that Google was considering an acquisition of HubSpot. While those talks fell through, it’s worth exploring why HubSpot made a compelling acquisition target for the search giant.
The reason has everything to do with the evolution of generative AIs like ChatGPT, the risk they pose to Google Search, and the opportunities they create for HubSpot.
Let’s start with the high level view:
Google’s most profitable products are first sit at the top of the funnel, acting as a source of traffic other companies’ websites.
This is true of Google’s core product (search), its social media property (YouTube), and its $237 billion dollar array of advertising products.
For the majority of B2B businesses investing in digital marketing, the middle of the funnel is some mix of:
- website, landing pages, lead-generating offers (ebooks, webinars, white papers, newsletters), product pages and sales pages
- a series of automated follow up emails and other forms of educational content
- an initial conversation with a junior sales rep to qualify the prospect, followed by many follow up conversations with more senior sales reps
- A demo, free trial, or freemium app + new user onboarding experience
A well-crafted middle of the funnel filters out less-than-ideal customers, educates ideal customers until they’re ready to buy, and prepares them to experience initial success once they do.
Generative AI (ChatGPT, Claude, Gemini) can already facilitate many of the tasks involved:
- Writing, designing, and optimizing websites
- Setting up landing pages and and drafting copy for marketing and sales emails
- Generating blog posts, imagery, and marketing videos
- Answering pre-sales questions and qualifying new prospects
- Visualizing and analyzing the data that these efforts produce
In theory, Google could leverage its own AI products, its massive advertising platform, and its productivity apps (like Google Docs/Sheets) to build a holistic solution for marketing and sales AI that automates all of these tasks.
The beginnings of this scenario is exactly what Google demonstrated in this section of the recent Cloud Next Keynote:
In practice, there are two essential pieces missing:
- The apps to deliver everything that happens AFTER a prospective customer clicks the ad (aka the middle and bottom of the funnel)
- The data set clarify which marketing, sales, and support activities generated the most revenue and created the most profitable customers
Without these, Google’s AI offerings can generate ad campaigns, Google Drive folders full of blog posts and marketing copy..and videos with synthetic actors and voices for days.
But it’s missing the product set necessary to integrate them into a complete system for marketing, sales, and customer support. See where this is going?
To offer a complete solution for AI marketing and sales, Google needs products that address the middle and bottom of the funnel.
In other words, it either needed HubSpot or the equivalent of HubSpot.
Given HubSpot’s unique position as an “all-in-one” CRM that includes tools for building websites, creating the equivalent of HubSpot would involve acquiring a last-generation “all-in-one” app like InfusionSoft or buying lots of different companies and welding their products around a shared customer database.
Google would need:
- a website and blog builder like Webflow, Wix or Squarespace
- a marketing automation app like ConvertKit or Customer.io
- A sales pipeline management app like Pipedrive
- a customer support app like Intercom or Zendesk
- a workflow builder like Zapier
- A whole lot of work to stitch them all into a coherent interface.
Google’s AI + an all-in-one CRM for every stage of the middle and bottom of the funnel could evolve into flywheel that spans every stage of the customer relationship…with the capabilities to analyze the results and optimize the performance of every piece over time.
This flywheel would scale to any small business, high growth company, and medium-sized enterprise that relies on digital and content marketing, high touch sales, and partnerships for distribution…in essentially any B2B niche.
Ah, but this assumes that Google’s position at the top of the funnel is secure.
Not so fast!
The emerging risk to Google Search
As the quality of answers and creative output from AI models improve, the relevance and utility of Google Search and the value of Search Ads faces a fair amount of risk.
How much risk? It would be foolish for someone without Nassim Taleb-level math and risk analysis skills to try to quantify it.
But it’s easy to imagine that over the rest of this decade, generative AIs will get better and better at answering questions for searches that suggest an intention to make a purchase:
- Searches about a clearly-defined problem or need: How do I generate leads for my plumbing business? What are family friendly vacations for a family of 5 with 3 young kids and a budget of $400/person?
- Searches for potential solutions: How do I choose a listing service for my plumbing business? How do I grow my plumbing business with social media? How do I find the best deal for all-inclusive-family vacations in the tropics?
- Searches about a specific product: What are the best alternatives to HomeAdvisor for plumbers who want to get more leads? What are less “touristy” alternatives to Club Med for family vacations in Costa Rica?
These types of searches map to classic “funnel” stages: Problem Aware > Solution Aware > Product Aware
And these three stages of the funnel are where Google Search ads make nearly all of their money.
If and when AIs advance to support up-to-date indexes of the internet (rather than searching Google or Bing and summarizing the results)…the use of Google Search for researching products and services may decline precipitously.
…along with the utility of placing ads next to those search results.
Google’s leadership is surely aware of this risk.
So if Google’s rumored interest in HubSpot is not just a made up story, it may have less to do with connecting the dots between its current ad products and the revenue-generating transactions that those ad products exist to support:
…And more to do with the urgent need to find the business model that comes after search advertising.
HubSpot’s Value to Google > Google’s Value to HubSpot
HubSpot has already begun to integrate generative AI across its entire set of products.
As of May 2024, its AI features can generate:
- Complete, editable websites built on its Content Management System
- Blog posts (including SEO-optimized titles, outlines, complete posts, and images)
- Reports and dashboards (e.g. how did each sales rep perform this quarter)?
- Prospecting emails
- Answers to support requests
CTO Dharmesh Shah recently outlined the roadmap for more.
Right now, these AI features are primarily time and labor savers: they offload repetitive tasks and empower a single person to produce a larger volume of completed work.
The generative AI tech that powers them is neither mature nor deeply integrated enough to reliably produce high-quality finished work without humans involved.
But it seems like only a matter of time before HubSpot AI can launch an entire HubSpot website with all the necessary pages, a blog with a collection of educational posts, a series of marketing videos or webinars, podcast episodes in the founders voice, and many sets of automated email campaigns from a single set of prompts.
At first, these will likely be first drafts that human professionals refine.
But at some point in the not-distant future, the whole thing is likely to be good enough to use as is. And with the customer data in HubSpot’s CRM, AI will be able to tailor all of the language and design elements for every specific audience a business serves.
This is the foundation of the AI-powered sales and marketing flywheel described above.
And it is HubSpot’s potential to become the ”hub” of that flywheel for small and medium sized enterprises and B2B startups that makes HubSpot’s opportunity as an independent company so interesting.
But there’s a piece missing: an offering that addresses the top of the funnel.
This brings us back to Clearbit, the customer-data enrichment product that HubSpot acquired for $150M in December of 2023.
In plain English, Clearbit takes a business email address (yourname@comapnyX.com) and has the data to connect the dots to the rest:
- Personal information (name, company name, title)
- LinkedIn Profile URL
- Twitter Handle
- GitHub username
- Current Companies
- Company IP address
- in some cases, previous companies and roles
It also offers features to apply this data in real time (like notifying the sales team when a prospect from XYZ company lands on a website or auto-completing a long lead-generation form with just an email address.)
Some clever founders (like Siqi Chen of Runway) use Zapier to run Clearbit’s data through OpenAI’s GPT4. Runway’s workflow qualifies every new prospect, sends a summary of qualified leads (and the reasons they’re a great fit) to Slack, and drafts personalized outreach emails for sales reps to review and send.
Before HubSpot acquired it, Clearbit also offered an advertising tool for building granularly-targeted B2B audiences and reaching them on Facebook. The pitch: pay 10x less what you’d pay on LinkedIn to reach the same audience on Facebook and Instagram.
So here’s where things get even more interesting:
The combination of HubSpot and Clearbit could evolve into the foundation of an AI-powered B2B advertising app.
Maybe also: a transformative new approach to B2B Go-to-Market Strategy
With data from across HubSpot’s CRM and Clearbit, HubSpot has potential insight into:
- Companies in the market for any of the products and services offered across HubSpot’s customer base
- Specific people and roles involved in the buying decision
- Buying processes for any companies served by HubSpot’s customers
They have this data at the level of a specific company and at the level of category (industry, segment, company size, etc)
In the near and medium term, the ethics and legality of enriching Clearbit’s data with HubSpot’s CRM data are questionable and poorly-defined (at best):
Also: Would any HubSpot customer be happy to learn the data they stored in HubSpot was used to inform a competitor’s ad campaigns?
But HubSpot doesn’t need to tread such murky waters to go in interesting directions: enriching their own customers’s data with Clearbit is more than enough to build an app for reaching B2B audiences across the major advertising platforms.
To start with, HubSpot could:
- Develop ad management tools like Shopify’s but for B2B, offering a user-friendly “Ads Hub” to launch and optimize campaigns that run across Facebook, LinkedIn, Google, (etc) straight from HubSpot’s interface.
- Build in AI to create specific ad copy, images, and videos, help refine the audience targeting, develop and run A/B tests, analyze the performance of each channel, and allocate budget accordingly.
- Build a recommendation / ad network for B2B (similar to Substack’s recommendations or ConvertKit’s Creator Network) that offers HubSpot customers and partners a means to earn revenue by recommending complimentary products and services to their customers.
While those may seem like generic features, HubSpot is well-positioned to connect the dots between the ad campaigns, the marketing and sales efforts, and the lifetime value of every customer.
Here’s what HubSpot CEO Yamini Rangan said about Clearbit on a recent earning’s call:
Phase one of integration was really to bring the Clearbit product into the hands of our installed-base customers. And, that is going well. It’s, again, early days. But, really the thing that we are excited about is natively bringing the company enrichment data, the intent data, and the contact enrichment data natively into HubSpot so it can power all use cases.
And, this is the way we see is the world operating. When you start thinking about a campaign, you already have the information about the companies that look like your customers, and then you can apply AI tools on top of that so that your campaigns can be supercharged. So, the combination of enriched data, AI, and a unified customer platform is really powerful, and we feel very good about where we are going in terms of that vision
For HubSpot, building advertising tools would be a logical next step towards the complete AI-powered marketing, sales, and customer support system we talked about earlier…the one that Google might like to acquire HubSpot to deliver…but that HubSpot could probably deliver on its own.
These features also offer the foundation of an even-more interesting and potentially-transformative opportunity to change the way B2B products get marketed and sold.
…and maybe even the upgrade the way new products get conceived, validated, and brought to market for the era of AI.
We’ll come back to this wilder speculation later.
First, let’s go deeper into why HubSpot may be more valuable to Google than Google is to HubSpot.
A tour of relevant regions of the digital ad landscape
The whole digital advertising ecosystem is beyond the scope of this essay.
For now, the key concept: in digital advertising, there are the “walled gardens” of Google Search, YouTube, TikTok, Instagram, LinkedIn, Twitter (etc) and “the rest of digital ad landscape.”
The walled gardens each have their own native advertising tools, formats, placements, and audience targeting criteria (Like search ads that target specific keywords, YouTube ads that target specific content types, Meta’s newsfeed and Story ads, etc)
The rest of the digital advertising landscape now extends far beyond banner ads on websites into mobile apps, streaming TV, highway billboards, terrestrial radio and podcast networks, and some of the animated signs you see in airports and transit systems.
For over a decade, a growing percentage of ads both inside and outside the walled gardens have been bought and sold “programmatically”–using apps with sophisticated algorithms to bid on ad inventory and place ads that reach specific audiences across the internet.
According to EMARKETER, spending on programmatic ads in the United States grew from $4.99 billion in 2013 to $135.72 billion in 2023: an increase of 2,719%.
These ad placements get bought and sold in auctions on ad exchanges: vast marketplaces of all of the ad space available outside the walled gardens at any given time.
Along with its walled gardens (Search, YouTube, Gmail, etc) Google operates one of the biggest digital ad marketplaces: the Google Display Network.
When it comes to all the ad inventory outside of the walled gardens, the Google Display Network is far from the only game in town .
Accordingly, sophisticated digital advertisers treat the Google Display Network as one possible option…when they can reach the same audience for a lower price, they often do.
The apps sophisticated advertisers use to accomplish this are called Demand Side Platforms (DSPs).
DSPs:
- Have visibility into available ad inventory across multiple media publishers, ad exchanges, and marketplaces
- Have the tech to bid for and buy digital ad slots in milliseconds (as web browsers load)
- Offer tools that let sophisticated advertisers build and optimize campaigns to reach their intended audiences across the digital ad landscape.
In the major DSPs, Google’s Display Network is one of many networks, and it competes for ad dollars with dozens of other vendors who can reach the same audiences.
Let’s bring this back to the HubSpot/Google Story:
As:
- Generative AIs become more effective than Google Search at delivering high quality, immediate answers to questions about products and services purchases
- Google Search results get noisier and noisier with AI-generated SEO content
- The ecosystem of ad-supported streaming video channels hooked up to Demand Side Platforms expands
- State regulators continue to treat the tech giants as political punching bags
The utility of Google’s most profitable “walled garden” ad product (search) is at risk.
While YouTube seems likely to remain a differentiated product, major streaming video services (Netflix and Amazon Prime) are already building out their ad platforms to get a slice of the video ad budget. The rest of Google’s Display Ad Network may be the largest and most technologically advanced, but it is not as differentiated as its walled garden products.
Meanwhile, HubSpot’s bundle of CRM apps empowers it to apply AI from any generative AI vendor (OpenAI, Anthropic, Google, etc) and choose the models most suitable for each task.
And with high-quality customer data in its own databases, HubSpot can offer ad tools for reaching granularly-defined B2B audiences across the walled garden ad platforms AND the rest of the digital ad landscape…
While building ad tools as sophisticated as a demand side platform would be outside of HubSpot’s core competence, it’s much easier to partner with an advanced ad tech vendor than it is to establish a unique position in “all-in-one CRM” for Small and Medium-sized enterprises.
(Also: as we’ll explore in a future piece, AI-facilitated software development may significantly lower the risk of exploring markets outside of a company’s original core competencies).
In other words: While HubSpot is well-positioned to build a B2B advertising app that addresses the top of the funnel…Google is not currently well-positioned to offer an “all-in-one” CRM that addresses the middle and bottom.
All of this may put the rumored acquisition in context.
If these rumors are true AND the bear case about HubSpot’s path to profitability is accurate, an acquisition at a 25-30% premium on HubSpot’s current stock price would be a great deal for HubSpot’s shareholders.
And it may be possible that Google’s vast resources, budget, and AI expertise would give HubSpot the runway and access to engineering talent and tech that it could not get any other way.
On the other hand, if this essay’s hypotheses about HubSpot’s opportunities with AI are accurate, HubSpot has only begun to tap into the value AI can add to its products.
Its potential to evolve into the hub of an AI-powered sales and marketing flywheel is more likely to be realized in an actual, working product if HubSpot remains free to implement the AI models best suited for the “job” of each aspect of marketing, sales, and customer support.
…And less likely if HubSpot gets entangled in the web of Google’s giant org structure.
And that opportunity may only be the lead-in to an even bigger, truly transformative possibility.
AI Matchmaking: The Logical Conclusion of Inbound Marketing
As generative AI matures, the entire way b2b products got marketed and sold seems likely to shift.
Today, the primary role of the marketing and sales functions is to filter out “poor fit” customers and maximize the probability that “great fit” customers find the product, choose to purchase it, and have the resources and support necessary to get the value they paid for.
This is the whole reason companies have websites, blogs, lead-generating offers like webinars, white papers, ebooks, build out automated email marketing campaigns, and invest in ads.
Facilitating this process also what effective, high-integrity sales people do every day.
On the buying side, there are often multiple people and roles involved:
- researching and comparing potential solutions to problems the business is facing
- championing a specific product
- project managing a pilot
- testing the product’s technical claims.
In many cases, there’s a whole buying committee to ensure that a major purchase and rollout of a new product is more likely to generate value than destroy it.
But more advanced generative AI has the potential to eliminate the need for a lot of that.
AI Matchmakers
In the worlds of romance, business recruiting, and professional networking, successful matchmakers have the knowledge and intuition to connect people to the best-suited partners, job opportunities, and professional contacts for their unique blend of life circumstances, needs, desires, and tastes.
To perform the task well, matchmakers must retain a degree of neutrality and awareness of the needs and goals of everyone involved: the match must be a great fit everyone, or it’s not a match.
As generative AI matures to the point where every business and person on the internet has their own AI agents with knowledge and context on their needs, desires, and circumstances, the whole way products and services get bought and sold could shift to “AI-facilitated matchmaking.”
Instead of searching the internet, reading reviews, and analyst reports while researching products to buy, the buyers of the future might instead have a conversation with their AI agents.
These AI agents would ask clarifying questions to determine what the buyer wants, why they want it, and the price ranges that makes sense.
They would already have background context on the buyers’ preferences, desires, needs, and values…as well as up-to-date and historical details of their businesses and the market landscape.
They also would not be trying to “sell” the buyer anything…only to find the ideal vendor and gather the information necessary to inform that choice.
The business owners and leaders of this future might have a similar conversation with AI agents about their products and services.
Those AI agents would not be trying to maximize sales from every possible source. Instead, they’d clarify the types of customers that the businesses’s offerings serve well and those they don’t…and give clear guidance on how to expand the range of possible customers or serve the existing ideal customer types more deeply.
In this near-futuristic scenario, the role of “marketing and sales” would be simplified to this:
- A businesses’ AI agent creates a “digital presence” that acts as useful and accurate “beacon” for the AI agents of their ideal customers.
- These AI agents ensure that buyers’ AI agents have the information necessary to support clear, well-informed, and accurate buying decisions.
- They “qualify” ideal customers and redirect customers that would be better served by another company’s offerings to that company digital presence.
- Once the buying decision is made, the AI agents from both buyers and sellers establish a rollout plan that takes into consideration as many of the factors and potential risks as possible and adapts in real time to new information and changing circumstances on both sides.
If this sounds like an “all-in-one” marketing, sales, and customer support system, but for AI agents talking to other AI agents rather than humans searching, scrolling, and having Zoom calls…
That’s because it is.
In a previous section, we imagined using the AI features on HubSpot’s roadmap to launch complete digital marketing and sales infrastructure from a short set of prompts: a website, a blog full of educational posts, marketing videos, automated email campaigns, chatbots for sales and customer support.
We speculated that HubSpot could also build its own ad tools for driving traffic to all of it.
In that scenario, HubSpot’s AI makes it radically more efficient to create infrastructure and content for humans marketing and selling to humans.
Evolving into infrastructure for “AI matchmaking” between B2B companies and the customers they are best-suited to serve would be a logical evolutionary next step.
It would represent the full actualization of HubSpot’s founding Big Idea:
Inbound Marketing in the age of generative AI: You create the products and services that life has made you perfectly suited to offer, and AI ensures that your ideal customers find you when those products and services are exactly what they need.
Realizing THAT possibility would turn HubSpot from an interesting business into a transformative company: an essential part of a much bigger story where the next generation of technologies empowers a larger and larger percentage of the human’s on Earth to realize their potential and take big steps towards living the life of their happiest dreams.
