Category Archives: AI

How To Navigate the Pipeline Crisis

Unlike many marketers, I’m not particularly prone to hyperbole, and thus “crisis” is not a word that I use lightly.  But I think saying “pipeline crisis” is warranted today when discussing what’s happening in marketing and is a key underlying cause for the broader malaise in SaaS growth

You don’t need to look far to find signs of a problem:

  • SaaS stocks, as measured by Bessemer’s Emerging Cloud Index are down 3.4% year to date.
  • Customer acquisition efficiency is down.  Earlier this year, median CAC payback periods hit 57 months, implying a staggering almost five years to recoup the cost of acquiring a dollar of net-new ARR.
  • Pipeline coverage ratios are running below their required targets.  The top reason for missing sales targets is insufficient pipeline coverage and Cloud Ratings shows stated coverage of 3.6x vs. target coverage of 4.1x.  (I can hear the cries of CROs everywhere saying, “please, just give me more at-bats!”)
  • Articles about the web traffic crisis are ubiquitous, from Rand Fishkin’s must-read posts on zero-click marketing to CJ Gustafson swimming outside his normal lane with a post entitled Google Zero.  The web is transitioning into a series of walled gardens and what’s left over is increasingly front-run both by Google search and, of course, answer engines such as ChatGPT, Perplexity, Claude, and Gemini.
  • Earlier this year, Andrew Chen put it bluntly:  Every Marketing Channel Sucks Right Now.

Add it all up and you can summarize this rather grim picture — as the Exit Five newsletter recently did — with Nothing Works Anymore.

I see this every day in my work with dozens of SaaS companies.  Because many companies are missing bookings targets by roughly the same percentage as they are missing pipeline coverage targets, I believe this is a pipeline crisis, and not a conversion rate crisis.

The struggle is real.  If you’re facing it, you are not alone.

Against this cacophony we hear a lot of talk about “brand vs. demand.”  The argument being that since demand generation programs are working less effectively, marketers should increasingly allocate dollars to brand programs.  It’s not a bad argument — in part because I believe that marketers over-rotated to highly measurable marketing during the go-go days — and thus a swing back to less directly measurable marketing is a good idea. 

(Aside:  I’d argue that marketers didn’t over-rotate on their own.  They got an assist from CEOs and CFOs who were only too eager to invest exclusively in marketing programs that delivered a clear short-term return and ignore the underlying complexity in B2B sales, effectively living-the-lie that is marketing attribution.  We don’t sell toothbrushes here, people.  Nobody goes to a tradeshow and buys a $250K enterprise solution — or even a $25K one — based on one interaction with one person.  But I digress.)

The question, of course, is what to do about it?

What Others Are Saying

A lot of smart people are weighing in, so I thought I’d provide a few links before sharing my own take.

  • Kyle Poyar wrote a great post called The 2025 State of B2B GTM Report.  (Subtitled “What’s Working in GTM?  Anything!?”)  My favorite part is the GTM Scorecard, a quadrant that maps channels by popularity and likely impact.  The underlying report is full of good ideas, GTM tool recommendations, and survey data.
  • The aforementioned Exit Five post, despite its title, is actually about what is working with answers derived from an informal poll of community members.
  • Scale recently published a State of GTM AI report which provides survey data on AI within GTM, focused largely on high-level use-cases and a two-phase adoption model.  (Jadedly, if we’re going to do less effective work, then let’s at least do it more efficiently.)
  • If your issues are more strategic, such as identifying and targeting sub-verticals, then you should read my friend Ian Howell’s book, Smart Conversations.

What Would Dave Do?

I’m going to build upon a popular comment I made on Kyle’s CAC payback period post.  Consider this a sister post to What To Do When You Need Pipeline in a Hurry, but this time not focused on the hurry, but on today’s environment.

Here’s what I would do:

  • Think holistically.  You might only be the CMO, but you need to look across all pipeline sources.  The job is to start quarters with sufficient coverage and notably not just to hit marketing pipegen goals.  If outbound is working, reallocate money to it.  If AEs can generate more pipeline (e.g., formal targets, more direct routing of inbound), then do it.
  • ABM.  Substitute across-the-board campaigns with targeted outreach on key accounts, leveraging both marketing and human channels (e.g., SDRs), both digital and dimensional assets (i.e., physical things like branded Moleskines), and intimate live events.  As an old CRO friend says, “if by ABM you mean us picking our customers as opposed to them picking us, then I am in favor.”
  • Events.  People are tired of working from home all day and champing at the bit to get out and press the flesh.  This includes major tradeshows, annual user conferences,  and roadshows all the way down to field-marketing dinners and sporting event boxes.
  • Get good at AEO.  It’s quickly replacing and more effective than search.  It’s also more winner-take-all.  There is plenty of content out there on how to do it and agencies eager to help.  Read these two articles for starters.
  • Leverage the CEO via social media (e.g., LinkedIn), podcast appearances, and speeches.  And LinkedIn doesn’t just mean a few posts, it means an overall strategy.
  • Use your AI message to put butts in seats.  We’re still in the stage where people are confused about AI and nothing puts butts in seats like confusion.  Do educational webinars, videos, and content.  Educate people but be sure to do it en masse.
  • Leverage AI tools and workflows.  Review Kyle’s report, particularly the part on the GTM tech stack.  Read Paul Stansik’s practical posts on AI, including how to avoid slop.
  • Build first-party audiences.  If you can no longer pay a reasonable amount to reach other people’s audiences, then you’re going to need to build your own.  While this is a slow burn, over time you’ll be happy you did it.  Build a Substack, a YouTube channel, a quality newsletter, or a podcast.
  • Leverage partners.  They can account for 20-30% of your pipeline and usually bring opportunities that close faster and with a higher conversion rate.  If you have a partner program, leverage it.  If you don’t, start building one.  It’s another slow burn, but you’ll be happy you did it.
  • Check your nurture tracksLong-term nurture is easily forgotten.  Measure recycled leads.  Report on your tracks.  Ensure you’ve built specific tracks for competitive loss and bad timing.  A/B test them, the flows, and the content.
  • Understand why you lose.  While I believe most companies have a coverage problem, not a conversion problem, I like to win anyway and if your conversion rates are below 20-25% you need to understand why.  Do quantitative win/loss via CRM reporting, listen to call recordings, and do win/loss interviews to understand what’s really going on.
  • Invest in customer success.  While I know this doesn’t help with pipeline coverage (except for expansion), always remember that the cost to backfill churn is CAC-ratio * lost-ARR.  Thus, if your CAC ratio is 2.0 and you lose $2M in ARR, it’s going to cost $4M to backfill it. The easiest – and most cost-effective — way to keep the ARR bucket rising is to limit leakage.
  • Join a community.  In times of change it’s important to have colleagues you can talk to, so I’d not only keep in close touch with existing peers, but join a marketing community like Exit Five to engage in shop talk.

Slides From My SaaS Metrics Palooza 2025 Session on Selling Work vs. Selling Software

Today, I presented at SaaS Metrics Palooza 2025 on the differences between selling work and selling software. I’d like to thank my metrics brother, Ray Rike, for inviting me to speak and I’d like to thank everyone who attended the session.

Topic covered include:

  • Defining outcomes
  • Contrasting outcomes vs. usage
  • The outcomes stack and intermediate vs. end outcomes
  • How a dating site would price based on outcomes vs. subscriptions
  • The basic trade-offs in selling subscriptions vs. outcomes
  • How to capture value created and share it between the vendor and customer
  • How selling outcomes can (radically) expand the total available market (TAM)
  • Jevon’s Paradox and what happens when we make things radically cheaper
  • Selling virtual humans vs. jobs-to-be-done
  • A long list of links to references for additional reading

You can download a PDF of the slides here. You should be able to see a recording of the session here. (Frankly, I’m not 100% sure that link will work, but you can try.) And I’ve embedded the slides below.

Smart Conversations by Ian Howells: A Must-Read Book on Where B2B Marketing Strategy Meets Generative AI

I first met Ian Howells in London long ago, as fellow footsoldiers in the early relational database wars. While you had to be pretty technical to do product marketing in those days, Ian was technical with a capital T, having just sprung from university with a PhD in distributed databases. We fought together on the losing side of the database wars [1], shared many of the same scars from the experience, learned many of the same lessons, and — I’m reasonably sure — both decided to aim our careers towards marketing to understand the dark and mysterious magic that was said to have been responsible for our misfortune [2].

I kept in loose touch with Ian over the years as he went to Documentum (content management) [3], SeeBeyond (supply chain), Alfresco (content management), and eventually to Intacct (accounting), later called Sage/Intacct after their subsequent acquisition by Sage.

So when I heard Ian wrote a book on how to use generative AI to improve marketing, I was intrigued. When I learned he was so excited about generative AI’s potential that he took a year off from work to dedicate all his time to the task, I was hooked. Whatever he produced, I was going to read it.

What he produced was a book called Smart Conversations, Revolutionizing B2B Marketing with the Generative AI Playbook. And in this post, I’ll share my conclusions based on a pretty in-depth reading of his book.

Here they are:

  • Anyone in B2B marketing with an interest in strategy should read this book.
  • This book isn’t what I expected. I feared the book might be full of prompts for generating content marketing (aka, AI slop), copy for marketing campaigns, presentations, or infographics.
  • Instead, Ian has produced an elegant work that teaches B2B marketing strategy while showing how to use generative AI to define and implement it. I’m not 100% sure what I was expecting, but this sure wasn’t it. It’s way, way better.
  • The book is both theoretical and applied. One page he’s explaining why you should target what I’d call sub-segments (that he calls micro-verticals). Five pages later he’s walking you through the prompts he uses to to build lists of them, right down to their NAICS codes.
  • On one page he’s talking about the definitions of ideal customer profiles (ICPs) and improved Geoffrey Moore positioning templates. A few pages later he’s got you in the prompts for getting ChatGPT to generate them. One minute he’s talking theoretically about the opportunities created by market discontinuities and, boom, several pages later, he’s back in the prompts showing you how to use ChatGPT to discover them.
  • What’s even more fun is how he shows what it used to take to do some of these exercises. Like building messaging by doing deep customer interviews, transcribing your notes, printing them, and then spreading them over a conference room for days so you can spot patterns. And then contrasting that to just how fast you can do it today.
  • This wonderful pattern repeats, through competitive analysis, all-in-one positioning, power messaging, and “wall of sound” campaigns [4]. Each time, the theory and then the ChatGPT practice.
  • Ian concludes with measurement. That section comes complete with a lesson on the benefits of becoming a market leader (that we both learned from the sting of Oracle’s lash), with lessons quite similar to what I describe in The Market Leader Play.

Congratulations to Ian on writing such a great book and sharing it with us. I’m glad you took the year off to write it! Now, every B2B marketing leader should go read it. Kindle version here.

Notes

[1] It’s not every day you find one of your company’s anti-competitor documents in the Computer History Museum!

[2] The company, by the way, was called Ingres. But since few have heard of Ingres today, I remind people they’ve almost certainly heard of its offspring: Postgres, which stood for Post-Ingres, an open source and extensible version of the system that achieved enormous popularity. I often say that “Postgres is corn” in the sense of The Omnivore’s Dilemma (i.e., it’s in everything) or quip that Postgres is Stonebraker’s revenge. While Larry Ellison made all the money, Stonebraker did win a Turing Award, create several new classes of database systems (e.g., column-oriented), and build Postgres which while ranking fourth on db-engines is generally acknowledged to have a higher market share than Oracle, in part due its open source heritage.

[3] And one of the original case studies in Geoffrey Moore’s classic, Crossing The Chasm.

[4] I’m not capable of typing the words “wall of sound” without referencing the Grateful Dead’s amazing and utterly impractical public address system. What Ian’s describing is what I call a backfire or surround-sound campaign, the goal being the economic buyer at your target can’t stop hearing about you from all sides. Regardless of the name, it’s a great idea, and a much more realistic goal on a limited budget than making “everyone” hear about you (e.g., super bowl ads).

My Fourth Appearance on AI and The Future of Work

This is a quick post to announce my latest appearance on Dan Turchin’s AI and The Future of Work podcast.

Dan, the founder/CEO of PeopleReign, has been doing AI since long before it was cool and, to give you an idea of how long he’s been podcasting about AI and the future of work, my appearance marks episode 324 of his podcast. He’s no Johnny-come-lately to the fascinating intersection of people and technology and his material is always worth a good listen.

In what’s become a tradition, I’m back on the show to talk about my 2025 predictions blog post. In the 43-minute episode we bounce around a lot, but cover these topics:

  • The evolution of search: answers, not links
  • LLM optimization, how to show up in LLM-generated answers
  • Why it’s dangerous to think you’re lost in a “sea of sameness” when it comes to product differentiation
  • Why branding isn’t the last bastion of differentiation
  • Why to track Rand Fishkin when it comes to the evolution of SEO to LLMO
  • Why general-purpose databases are generally good at absorbing special-purpose databases — but not always
  • Does Europe’s tendency to greater regulate have any hidden benefits?
  • The Robin Williams quote about Canada: “it’s like living in the apartment above a meth lab.”
  • How America-first VCs will likely shoot their feet off with European companies and entrepreneurs
  • Why I predicted that LinkedIn will likely follow the path to enshittification by following engagement as their north star

I’ll conclude by saying that the Future of Work has become one of my favorite topics. It started with my Clubhouse room (remember Clubhouse?) with Thomas Otter. That led to some ongoing collaboration with Thomas when he moved to become a partner at Acadian Ventures (which, by the way, is an investor in PeopleReign). That also eventually led, through introductions to the founders, to my joining the board of TechWolf, where I’m now learning about redesigning work, people-centric data platforms, and the skills-based organization. It’s a fascinating area, particularly here at the dawn of mainstream AI, and one that affects all of us.

Thanks again Dan for having me on the show and for a great conversation about the predictions and a whole lot more.

Why I’m Joining the Board of TechWolf

I’m pleased to announce that I’ve joined the board of Techwolf, a Belgian HR tech company backed by a slew of top venture capital investors including Harry Stebbings’ 20VC, Future of Work boutique Acadian Ventures, vertically-focused SemperVirens, and European-focused Felix Capital, Notion, and Stride.vc.  They also have world-class strategic investors including SAP, ServiceNow, and Workday.

I love when a process works.  This started with an introduction from a trusted friend and category expert, Thomas Otter.  I met with the founders now and again over the years, via the odd Zoom or a coffee on California Avenue when they were in town.  I like this go-slow approach because you get to know the team and the company.  You watch them grow.  You stay in touch.  And then one day an opportunity to work together more formally appears.

Now, let’s talk about what I like about Techwolf:

  • The founders, Andreas, Jeroen, and Mikaël.  Independent directors (known in Europe as non-executive directors) are more about coaching than governance.  Thus, you need great chemistry with the founders.  Your skills need to complement theirs.  And they have to want to learn from you. 
  • The story.  Three computer scientists meet in college, win a hackathon together, found a company for recruiting, realize it doesn’t work, then pivot to a successful strategy around skills management.  C’mon.  Goosebumps.  I love every element of it.
  • The space.  The skills-based organization is a powerful and transformative vision for the future of work.  It’s one that takes technology to implement.  And it’s a great use-case for AI, starting with the problem of building an inventory of skills for the people you have already — before hiring hundreds or thousands of new ones.  It’s a win/win vision because it means companies can do more with less all while providing employees with better growth paths and more stimulating work
  • The validation.  More than my opinion, I was impressed that experts like Jason Corsello (former head of corpdev at Cornerstone), Thomas Otter (former head of product at SuccessFactors, and former Gartner RVP covering the space), and Andy Leaver (former head of EMEA at Workday) all seemed to love the idea, too.  Not to mention the implied endorsements of SAP, ServiceNow, and Workday.
  • The data-centric approach.  Rather than building a classic app that simply links a UI to a database (and leaves the heavy lifting to someone else), the founders cut straight to heart of the problem.  Skills data is a data problem.  And the best data doesn’t live in HR systems; it lives in operational systems and external data sources.  Solve that and the rest is somewhat trivial by comparison.
  • The timing.  I believe this company is in precisely the right place at exactly the right time.  The skills-based organization is a white hot trend in HR and you need technology like TechWolf’s to realize it.
  • The board and investors.  They’ve built a great team here to support them on their mission.  And raised over $55M to pursue it.
  • The fit.  Ever since I moved to Paris to work at Business Objects, I’ve been working with European companies on growth strategies and US expansion.  Thanks to my operating experience in Europe, my board experience at companies like Nuxeo, and my EIR work at Balderton, I feel pretty qualified to help with this sometimes thorny problem.

Thanks to Thomas for introducing us, and thanks to Andreas, Jeroen, and Mikaël, for welcoming me onto the team.