Latest / Uphoff on Media Podcast / Before You Flatten the Org Chart, Make Sure You Can Measure What AI Delivers

Transcript

Before You Flatten the Org Chart, Make Sure You Can Measure What AI Delivers · Uphoff on Media Podcast

• 0:01
  SPEAKER_00: Hey, it's Tony.
• 0:02
  SPEAKER_00: Plenty of people want to call this the peak of the AI hype cycle.
• 0:06
  SPEAKER_00: Block's numbers say otherwise.
• 0:09
  SPEAKER_00: In February, Block cut about 40% of its staff.
• 0:12
  SPEAKER_00: Eight months later, it's running at roughly $2 million of gross profit per employee.
• 0:17
  SPEAKER_00: That's double where it started the year.
• 0:20
  SPEAKER_00: And it's raised its guidance twice.
• 0:22
  SPEAKER_00: So the scale is real.
• 0:23
  SPEAKER_00: That's not the problem.
• 0:24
  SPEAKER_00: The problem is what comes next.
• 0:26
  SPEAKER_00: Companies everywhere are flattening their org charts on the strength
• 0:29
  SPEAKER_00: of productivity they're hoping for, not productivity they can measure.
• 0:33
  SPEAKER_00: Block could do this because it runs on the most honest signal there is, money.
• 0:38
  SPEAKER_00: Most B2B companies would be reorganizing around opens, clicks, and lead scores.
• 0:44
  SPEAKER_00: So today,
• 0:44
  SPEAKER_00: which signals you can actually trust,
• 0:47
  SPEAKER_00: why revenue per employee just became the number buyer's price,
• 0:51
  SPEAKER_00: and how do you know you're ready before you touch the org chart?
• 0:55
  SPEAKER_00: Here's how to tell the difference.
• 0:57
  SPEAKER_00: This episode is titled,
• 0:59
  SPEAKER_00: Before You Flatten the Org Chart, Make Sure You Can Measure What AI Delivers.
• 1:03
  SPEAKER_00: Okay, let's get to work.
• 1:07
  SPEAKER_00: AI is delivering on the scale promise.
• 1:09
  SPEAKER_00: Block double gross profit per employee by rebuilding around honest data.
• 1:13
  SPEAKER_00: Most B2B companies don't have the signals to try it yet.
• 1:18
  SPEAKER_00: I'm a fan of radical thinking.
• 1:20
  SPEAKER_00: It forces you to test assumptions you didn't know you were making.
• 1:23
  SPEAKER_00: You don't always have to agree with it.
• 1:25
  SPEAKER_00: Agreement is not really the point.
• 1:27
  SPEAKER_00: The point is whether it makes you think.
• 1:30
  SPEAKER_00: Jack Dorsey and Roloff Botha's essay, From Hierarchy to Intelligence, did that to me.
• 1:36
  SPEAKER_00: It lays out how Block is rebuilding itself as a company organized
• 1:39
  SPEAKER_00: as intelligence rather than hierarchy.
• 1:42
  SPEAKER_00: And it came a month after Block cut its workforce by 40%.
• 1:46
  SPEAKER_00: Most of what gets written about AI and organizations stops at co-pilots,
• 1:51
  SPEAKER_00: a smarter assistant bolted onto the same old structure.
• 1:55
  SPEAKER_00: They go further and ask whether the organization itself still needs to exist
• 2:00
  SPEAKER_00: in its current form.
• 2:01
  SPEAKER_00: It pushed me past the easy cliche that AI just means fewer middle managers.
• 2:06
  SPEAKER_00: Here's where that thinking led.
• 2:10
  SPEAKER_00: The oldest constraint in business.
• 2:12
  SPEAKER_00: Dorsey and Botha traced 2,000 years of the same fix.
• 2:16
  SPEAKER_00: The Roman Legion built command around eight soldiers to a tent.
• 2:20
  SPEAKER_00: The Prussian general staff existed to route information up and down the chain.
• 2:25
  SPEAKER_00: Daniel McCollum drew the first org chart in the 1850s to keep a 500-mile railroad
• 2:30
  SPEAKER_00: from killing people.
• 2:32
  SPEAKER_00: Every version since, from Taylor's Pyramid to The Matrix, worked around one constraint.
• 2:37
  SPEAKER_00: A human brain can only hold so much context,
• 2:40
  SPEAKER_00: so organizations add layers and layers, slow down information.
• 2:45
  SPEAKER_00: Spotify, Zappos, and Valve tried to break the pattern.
• 2:49
  SPEAKER_00: All three scaled back into hierarchy or hit a ceiling.
• 2:52
  SPEAKER_00: Dorsey and Botha's claim is that AI removes the constraint, not the org chart, around it.
• 2:59
  SPEAKER_00: Okay, what Block actually built?
• 3:01
  SPEAKER_00: Block is building a company world model,
• 3:04
  SPEAKER_00: what the business knows about its own operations,
• 3:07
  SPEAKER_00: and a customer world model,
• 3:08
  SPEAKER_00: what it knows about the people transacting with it,
• 3:11
  SPEAKER_00: and using both to compose solutions directly without routing decisions
• 3:15
  SPEAKER_00: through management.
• 3:17
  SPEAKER_00: What's left are three roles,
• 3:19
  SPEAKER_00: individual contributors who build,
• 3:21
  SPEAKER_00: DRIs who own outcomes across whatever teams a problem touches,
• 3:26
  SPEAKER_00: and player coaches who develop people instead of running status meetings.
• 3:29
  SPEAKER_00: No middle layer required.
• 3:32
  SPEAKER_00: Money is the one signal nobody fakes.
• 3:35
  SPEAKER_00: Here's the sentence that matters most in the whole piece.
• 3:39
  SPEAKER_00: Quote, money is the most honest signal in the world, end quote.
• 3:43
  SPEAKER_00: People lie on surveys,
• 3:45
  SPEAKER_00: they abandon carts, they ignore ads, but when they spend, that's the truth.
• 3:49
  SPEAKER_00: Block can build a company as intelligence because it sits on both sides of millions
• 3:54
  SPEAKER_00: of transactions every day through Square and Cash App.
• 3:58
  SPEAKER_00: That gives it a signal almost nobody else has, honest, at scale, compounding.
• 4:04
  SPEAKER_00: Where this breaks for the rest of us,
• 4:06
  SPEAKER_00: that's also where the model runs into trouble for most of the people reading this.
• 4:11
  SPEAKER_00: Very few B2B information marketing or technology companies have anything close to
• 4:15
  SPEAKER_00: a money grade signal.
• 4:16
  SPEAKER_00: What they have instead is engagement data, opens,
• 4:19
  SPEAKER_00: clicks, content consumption, lead scores, intent signals bought from a vendor.
• 4:25
  SPEAKER_00: Look,
• 4:26
  SPEAKER_00: I've run enough pipeline reviews to know how much of that gets gamed on purpose by
• 4:30
  SPEAKER_00: the same marketing teams reporting on it.
• 4:32
  SPEAKER_00: The essay's point about people lying on surveys is more true of modern B2B marketing
• 4:37
  SPEAKER_00: than its authors probably intended.
• 4:40
  SPEAKER_00: A lead score isn't a fact about someone's business,
• 4:43
  SPEAKER_00: it's a fact about how well your content ranks.
• 4:46
  SPEAKER_00: Most B2B signal sits at the wrong end of the spectrum.
• 4:49
  SPEAKER_00: That gap is the actual work before anyone in B2B tries to run blocks playbook.
• 4:55
  SPEAKER_00: Think of it as a spectrum, honest signal at one end and performative signal on the other.
• 5:00
  SPEAKER_00: Self-reported data and form fills sit at the weekend, cheap to produce, cheap to fake.
• 5:07
  SPEAKER_00: Engagement metrics sit a notch above.
• 5:09
  SPEAKER_00: Harder to fake outright, but easy to optimize for the metric instead of the buyer.
• 5:14
  SPEAKER_00: Usage, depth,
• 5:15
  SPEAKER_00: and renewal behavior sit higher still,
• 5:17
  SPEAKER_00: but because pretending cost the customer something real.
• 5:21
  SPEAKER_00: Contract value, expansion, revenue, and churn sit closest to money itself.
• 5:26
  SPEAKER_00: The businesses already living near that end of the spectrum are
• 5:30
  SPEAKER_00: the ones actually positioned to build a world model worth trusting.
• 5:34
  SPEAKER_00: Everyone else is redesigning an org chart around a signal that isn't honest yet.
• 5:40
  SPEAKER_00: This isn't just Silicon Valley talking.
• 5:42
  SPEAKER_00: It'd be easy to dismiss this as some crazy idea from the tech elite.
• 5:46
  SPEAKER_00: Don't.
• 5:47
  SPEAKER_00: The data says otherwise, and it's not subtle.
• 5:49
  SPEAKER_00: The number of managers and executives at public companies fell 6.1% between May
• 5:55
  SPEAKER_00: of 2022 and May of 2025, according to LiveData Technologies.
• 6:01
  SPEAKER_00: Gardner predicted in October 2024 that through this year,
• 6:05
  SPEAKER_00: one in five organizations would use AI to flatten their structure,
• 6:09
  SPEAKER_00: eliminating more than half of current middle management positions.
• 6:13
  SPEAKER_00: We're in the last quarter of that window.
• 6:16
  SPEAKER_00: I've been the skeptic on AI-attributed layoffs.
• 6:19
  SPEAKER_00: In June,
• 6:20
  SPEAKER_00: I called most of it a misattribution problem,
• 6:22
  SPEAKER_00: pandemic correction and PE margin engineering relabeled as AI.
• 6:26
  SPEAKER_00: That's still true of last year's numbers,
• 6:29
  SPEAKER_00: but manager headcount specifically, not total headcount, is the number that's moved.
• 6:34
  SPEAKER_00: Nobody announced that.
• 6:35
  SPEAKER_00: They just stopped backfilling the role.
• 6:37
  SPEAKER_00: This isn't theoretical, and it isn't just block.
• 6:41
  SPEAKER_00: GitLab and Coinbase both cut management layers this year and pointed to AI when they did.
• 6:46
  SPEAKER_00: CloudFair's cuts reached middle management too.
• 6:49
  SPEAKER_00: Google said last year it eliminated 35% of its managers overseeing small teams.
• 6:55
  SPEAKER_00: Intuit, Meta, Cisco,
• 6:57
  SPEAKER_00: Atlassian,
• 6:57
  SPEAKER_00: Salesforce,
• 6:58
  SPEAKER_00: and Oracle all paired 2026 layoffs with AI investment or efficiency claims.
• 7:04
  SPEAKER_00: Block didn't start this.
• 7:05
  SPEAKER_00: It's just the company doing it with a strong signal This is the strongest claim to
• 7:10
  SPEAKER_00: a signal that actually justifies it.
• 7:14
  SPEAKER_00: Revenue per employee is how the market now keeps score.
• 7:17
  SPEAKER_00: Blocks reorg came with a scorecard that isn't the org chart.
• 7:20
  SPEAKER_00: Runblocks own numbers.
• 7:22
  SPEAKER_00: 2025, 10.36 billion of gross profit with 10,205 people, about a million per person.
• 7:32
  SPEAKER_00: $12.2 billion for 2026,
• 7:34
  SPEAKER_00: with just under 6,000 people, the stock jumped more than 20% after hours.
• 7:40
  SPEAKER_00: Now, guidance raised twice,
• 7:41
  SPEAKER_00: most recently in August to $12.51 billion
• 7:45
  SPEAKER_00: after second quarter gross profit grew 25%,
• 7:50
  SPEAKER_00: about $2.1 million per person, double where it started the year.
• 7:55
  SPEAKER_00: The market wasn't paying for the layoff, it was paying for the output per head.
• 8:00
  SPEAKER_00: Then, a vanity stat.
• 8:02
  SPEAKER_00: For most of my career, revenue per employee was a number nobody built a valuation around.
• 8:06
  SPEAKER_00: A 2013 analysis of public software companies found it had almost no relationship
• 8:12
  SPEAKER_00: to market value.
• 8:13
  SPEAKER_00: Hiring waves, sales ramp times, and company age moved it more than strategy did.
• 8:18
  SPEAKER_00: Now, the cleanest read on scale.
• 8:21
  SPEAKER_00: AI changed what the number measures.
• 8:22
  SPEAKER_00: It shows whether technology or headcount is doing the work.
• 8:25
  SPEAKER_00: And that's the question every buyer and investor is asking.
• 8:29
  SPEAKER_00: Here's where the bars sit.
• 8:32
  SPEAKER_00: Marketing agency average, about 163,000 per employee.
• 8:36
  SPEAKER_00: Median public software company, about 395,000.
• 8:39
  SPEAKER_00: Median AI native software company with more than 100 million of ARR, 655,000.
• 8:47
  SPEAKER_00: Top quartile of that group, $890,000.
• 8:52
  SPEAKER_00: Why it's hard to fake?
• 8:53
  SPEAKER_00: Anyone can say AI made their team more productive.
• 8:56
  SPEAKER_00: Revenue per employee shows whether the productivity reached the P&L.
• 9:00
  SPEAKER_00: That's why I added it to the Uphoff Value Index this quarter as a driver
• 9:04
  SPEAKER_00: in all eight categories that we benchmark.
• 9:07
  SPEAKER_00: A services firm approaching the software standard earns up to a full turn of EBITDA.
• 9:12
  SPEAKER_00: Two agencies with identical P&Ls,
• 9:14
  SPEAKER_00: one running with 68 people and one with 40,
• 9:17
  SPEAKER_00: now land nearly a million dollars apart in value.
• 9:21
  SPEAKER_00: One caution from the operator's chair, revenue per employee can be gamed too.
• 9:26
  SPEAKER_00: Push work to contractors or offshore partners and count only full-time staff and
• 9:30
  SPEAKER_00: the number looks better than the business is.
• 9:33
  SPEAKER_00: Expect buyers to start asking for revenue per dollar of total labor costs, not just per head.
• 9:39
  SPEAKER_00: Measured honestly,
• 9:40
  SPEAKER_00: it's the closest thing most B2B companies have to a money-grade signal
• 9:44
  SPEAKER_00: about themselves.
• 9:47
  SPEAKER_00: Marketing and sales break first.
• 9:49
  SPEAKER_00: If you're running marketing or sales, none of this is abstract.
• 9:52
  SPEAKER_00: These are the two functions most exposed, and the exposure isn't only about headcount.
• 9:57
  SPEAKER_00: It's about who you hire,
• 9:58
  SPEAKER_00: how you coach them,
• 9:59
  SPEAKER_00: and what good looks like once the AI is closer to the honest signal
• 10:03
  SPEAKER_00: than your team is.
• 10:04
  SPEAKER_00: Marketing.
• 10:05
  SPEAKER_00: The funnel was never honest.
• 10:07
  SPEAKER_00: Marketing has been running on performative signal for years and the funnel proves it.
• 10:11
  SPEAKER_00: Forrester has found that fewer than 1% of leads make it from the top of the funnel
• 10:16
  SPEAKER_00: to a closed deal.
• 10:17
  SPEAKER_00: The industry knows the model is broken and vendors are already pitching a replacement.
• 10:21
  SPEAKER_00: Docket calls it the agent qualified lead,
• 10:24
  SPEAKER_00: a lead qualified by an AI agent in a live buyer conversation,
• 10:28
  SPEAKER_00: capturing real intent instead of inferring it from a click or a form fill.
• 10:32
  SPEAKER_00: Hiring for what it That looks nothing like hiring for lead gen volume.
• 10:39
  SPEAKER_00: Coaching it looks nothing like reviewing MQL counts on a Monday pipeline meeting.
• 10:45
  SPEAKER_00: The SDRs who survive aren't dialing more, they're coaching AI performance.
• 10:50
  SPEAKER_00: Sales, the job is splitting in two.
• 10:52
  SPEAKER_00: It's happening right underneath your comp plan.
• 10:55
  SPEAKER_00: Regi.ai's analysis of LinkedIn data found full-cycle AE seats are up 10.8%,
• 11:02
  SPEAKER_00: while SDR headcount fell 22.7.
• 11:05
  SPEAKER_00: Two-thirds of teams cut SDRs.
• 11:08
  SPEAKER_00: Fewer than one in five added them.
• 11:12
  SPEAKER_00: The SDRs who stayed are running AI on the accounts that matter, not working the phones.
• 11:16
  SPEAKER_00: That's a different hire,
• 11:18
  SPEAKER_00: managed a different way,
• 11:19
  SPEAKER_00: measured against a different definition of quote of progress.
• 11:23
  SPEAKER_00: The playbooks most of us built our marketing and sales organizations on assumed
• 11:27
  SPEAKER_00: a human was doing the qualifying, the outreach, and the judgment calls.
• 11:31
  SPEAKER_00: Agenic AI is already doing large parts of all three.
• 11:34
  SPEAKER_00: Keep running the old playbook and you'll be coaching people to compete with
• 11:37
  SPEAKER_00: a system that's already faster and can't be gamed.
• 11:42
  SPEAKER_00: The real work comes before the reorg.
• 11:45
  SPEAKER_00: That's the mistake I expect to watch a lot of companies make over
• 11:48
  SPEAKER_00: the next two years,
• 11:49
  SPEAKER_00: skipping straight to flattening the org before doing the harder,
• 11:53
  SPEAKER_00: less glamorous work of finding or building a signal that can't be gamed.
• 11:57
  SPEAKER_00: Speed is the advantage Dorsey and Botha are chasing, and they're right that it compounds.
• 12:02
  SPEAKER_00: It compounds in the wrong direction just as fast when the thing moving fast is built
• 12:07
  SPEAKER_00: on noise, not signal.
• 12:09
  SPEAKER_00: Before anyone touches their org chart, they should answer one question honestly.
• 12:14
  SPEAKER_00: What's the one number in this business that nobody can fake?
• 12:18
  SPEAKER_00: If there's a real answer,
• 12:20
  SPEAKER_00: the reorg is worth having and revenue per employee will tell you within a year
• 12:24
  SPEAKER_00: whether it worked.
• 12:25
  SPEAKER_00: If there isn't, that's the actual project and it has nothing to do with headcount.
• 12:33
  SPEAKER_00: Okay, that's the episode.
• 12:35
  SPEAKER_00: If you want to put this to work,
• 12:36
  SPEAKER_00: two Uphoff advisory frameworks are in the post on Substack linked in the show notes.
• 12:42
  SPEAKER_00: The first is the signal to honesty spectrum.
• 12:45
  SPEAKER_00: Take every number your leadership team reviews each week and place it on the spectrum.
• 12:49
  SPEAKER_00: If most of them sit at the performative end, that's your first project.
• 12:53
  SPEAKER_00: The second is the revenue per employee ladder.
• 12:56
  SPEAKER_00: Find where your business sits today,
• 12:58
  SPEAKER_00: then decide where it needs to be before you touch the org chart, not after.
• 13:02
  SPEAKER_00: And if you found this useful,
• 13:03
  SPEAKER_00: please do one thing,
• 13:04
  SPEAKER_00: follow Uphoff on Media on Apple, Podcast or Spotify and leave a rating.
• 13:08
  SPEAKER_00: It's the single best way to help another operator find it.
• 13:12
  SPEAKER_00: I'm Tony Uphoff, back at it next week.