The R&D Metric Mark Hurd and HP Got Wrong

Twenty years. Nearly one thousand episodes on this show. And starting today, we're going to try something a little different this season.

Season 21 is about the decisions that actually determine whether innovation lives or dies inside any organization. The real calls. Not the fluff stuff we read in academic textbooks. I want to actually put you in the rooms where these decisions are happening. What went right. What went wrong.

My objective is to expose you to the patterns in innovation decisions so that you can recognize them. Recognize them in yourself, in the people you need to influence, long before you step into any landmines.

So let's get into it.

.. or listen to the podcast

The Encounter on the Top Floor of Building 25

Making generational decisions on innovation investment can be a make-or-break moment. What I refer to as a CLM, a "Career Limiting Move". In my case, it started with a chance conversation with Mark Hurd, HP's CEO.

Let me take you back to 2005. HP headquarters is on Page Mill Road in Palo Alto, referred to internally as Building 25. The top floor is where all the executive offices are. That's where Mark's office was.

I was up there doing some meetings and got snagged by Mark. Now, Mark had a reputation. He was a big numbers guy. He believed in what he called extreme benchmarking. You tore into your competitors' numbers. You knew your own numbers in and out.

Others had warned me about this. He had a famous quote that everybody shared: 

"Stare at the numbers long enough, and they will eventually confess." ~ Mark Hurd

Mark believed you could not lead a critical role at HP if you did not know your numbers cold, inside and out. Didn't matter whether it was sales, CTO, a function, or a division. It didn't matter. And Mark tested everyone on the leadership team. Not just the leadership team. He would randomly stop employees and ask them for their numbers based on what group they worked in. It was non-stop. It was constant. To where support staff was literally constantly preparing briefing books for managers, VPs, leaders, just in case they got nabbed by Mark.

In my case, I happened to be walking past his office. Mark waved me in. I sat down, and he immediately started drilling me on the CTO numbers.

The number he focused on was R&D as a percentage of revenue.

The Broken Benchmark: R&D as a Percentage of Revenue

Now, if you've been a regular listener of this show, you know my opinion of that metric. R&D as a percentage of revenue is a meaningless number.² It is absolutely meaningless. But every public company CEO at an innovation-dependent company, all the tech companies, AI companies, even automotive, they live by this number. It's a number that Wall Street looks at. You have to report it as part of your quarterlies, and from there it's simple math.

When Mark grilled me, he was focused specifically on the PC group at HP. HP's number at the time for the PC group was about one and a half percent R&D as a percentage of the PC group's revenue. Acer, which was a key competitor, was at 0.8%. Less than one percent. Roughly half of HP's number.

Apple was at four percent.

Mark's question, and he was really pounding on this, was: How do we get our ratios in line with Acer? Basically, he was saying: how do we cut costs so that our R&D expense as a percentage of revenue equals Acer at 0.8%?

This is exactly the problem with choosing the wrong metric.

Now I'm going to quote somebody who I think was probably one of the most insightful leaders in the business world. Charlie Munger. If you've ever watched any of his talks, he had a really strong opinion on certain metrics. Specifically EBITDA, earnings before interest, taxes, depreciation and amortization. Charlie referred to EBITDA as BS earnings. It was a metric Wall Street swore by, and Munger said it hid more than it revealed. His exact words: "Every time you see the word EBITDA, just substitute the word 'bullshit' earnings."

R&D as a percentage of revenue is the same problem in a different disguise. It's the metric that makes every company look like it's investing when all it's doing is spending.

Mark was using a broken instrument to make a generational decision. If you make decisions based on R&D as a percentage of revenue, and then you do comparisons like "let's make our numbers look like Acer," what you are actually deciding to do is cut your R&D. That is generational. You will destroy a company's innovation capability over the next ten to twenty years before you can even have a hope of rebuilding it.

"We Are Not Apple and We Never Will Be"

I looked at him and said: Why aren't we raising our R&D spend to match Apple?

Mark didn't hesitate. He said: "We are not Apple and we never will be."

I took offense at that. I was offended that he wouldn't even contemplate it. And I pushed back. I pushed back hard. I argued we could be Apple in areas where we had genuine advantage.

Here's one example. Go back to September 2004, about a year before my meeting with Mark. Carly Fiorina was still CEO. Carly had just handed Steve Jobs access to the retail shelf space HP spent thirty years building.

At that time, HP controlled about nine, nine and a half percent of all retail shelf space for consumer electronics, the largest single entity holding in that category. Where did all that come from? It traces back to the calculator days in the 1970s. Those relationships, those stocking slots, that footprint: HP had spent three decades building that access.

Apple was launching the iPod. It had no retail distribution in consumer electronics. None. And rather than HP taking advantage of that for itself, it actually opened the door and allowed Apple to come in. That is how the iPod got its traction. It bought Apple the time to build out its own retail strategy, which is ultimately what allowed Apple to be where it is today.

That wasn't an accident of history. That was HP giving away a structural competitive asset.

When I tried to push back on Mark, saying we could be better with the right investment, it didn't land. Mark viewed the PC business as a commodity. And if it's a commodity, you manage expenses. You don't invest in capabilities.

Monthly Arguments and the Search for Better Metrics

There was no decision made that day. But something shifted in me.

That was the first of many monthly arguments I had with Mark. And they were non-stop.

What it drove me to do was start looking for better metrics. We had something most companies don't have: HP's complete financial history going all the way back to the 1940s. I had access to the numbers, division by division, for one of the founding companies of Silicon Valley.

We were getting traction. I was actually getting Mark to align. I was getting the HP board to align. And then what happens? Mark gets removed as CEO and Leo comes in. Then Meg kicked Leo out and she took over. Then the split of HP into two companies.

Acer today? Still roughly 0.9% of revenue in R&D. Twenty years later, almost exactly where Mark wanted HP to get to.

What I Would Do Differently: Right Argument, Wrong Language

If I'm being honest about what I would do differently, I had the right argument. I had the wrong language.

The job wasn't to prove Mark wrong. Nobody changes their mind when they're being told they're wrong. I needed to stop speaking CTO and start speaking CEO. Meet him where he was. Make the case in the language of margin, risk, competitive position, the language he already trusted.

But that language didn't exist when it came to R&D and innovation. That's the reason I spent the rest of my career building something better.

And that is what this season is about.

What Comes Next: The Metrics That Tell the Truth

That conversation with Mark sent me looking. If R&D as a percentage of revenue was the wrong metric, and I believe to my core that it was, and is, then what's the right one?

We went back through HP's own numbers. We back-cast all the way to the 1940s, looking at the numbers by division, by the overall organization. And then something unexpected happened.

The archive team at HP gave me access to something nobody had looked at in decades: Bill Hewlett and Dave Packard's original notebooks.

What I found in there pointed me somewhere nobody had thought to look.

In the next episode, we're going to talk about the metrics that actually tell the truth when it comes to R&D and innovation.

If this episode gave you some insights, shifted something, share it with somebody who you think needs to hear it. Particularly if you're trying to fight senior leaders around R&D investment. And in the comments below, tell me:

What's that one benchmark that you are required to hit, and yet you've never questioned? Is it the right benchmark? Have you really looked at it?

Endnotes

1. "extreme benchmarking": Mark Hurd joined HP as CEO and President in April 2005 following Carly Fiorina's departure. Fortune reported that his "obsession with numbers permeates HP" and that he preferred to "talk in numbers than words," pointing to operating margins he drove from 4% in 2005 to 8.8% by 2008. "Mark Hurd's Moment," Fortune, February 2009, https://fortune.com/longform/mark-hurds-moment/. See also "Mark Hurd," Wikipedia, https://en.wikipedia.org/wiki/Mark_Hurd.
2. "a meaningless number": R&D as a percentage of revenue measures input, not output, and has no consistent correlation with innovation performance or long-term shareholder value. PwC's Strategy& Global Innovation 1000 study, conducted annually since 2005, has repeatedly confirmed there is no statistically significant relationship between R&D spending as a percentage of revenue and revenue growth, profit growth, or total shareholder return. Barry Jaruzelski and Kevin Dehoff, "Smart Spenders: The Global Innovation 1000," Strategy+Business, Winter 2006, https://www.strategy-business.com/article/06405. See also PwC Strategy&, Global Innovation 1000 Study, https://www.strategyand.pwc.com/gx/en/insights/innovation1000.html.
3. "you report your R&D spend": U.S. public companies are required under GAAP to expense and disclose R&D costs as incurred, making R&D as a percentage of revenue trivially calculable from any quarterly or annual SEC filing. For HP's actual reported figures by year, see Erik Sherman's compilation from HP annual reports: "Under Mark Hurd, HP Innovation Nose-Dived," CBS News, December 20, 2010, https://www.cbsnews.com/news/under-mark-hurd-hp-innovation-nose-dived/. For the GAAP standard itself, see SEC Staff Accounting Bulletin guidance at https://www.sec.gov/interps/account/sabcodet13.htm.
4. "HP's number at the time for the PC group was approximately 1.5%": HP did not break out R&D expenditure by division in its public filings. CBS News compiled HP's total company R&D from annual reports, showing 4.0% of revenue in FY2005, declining to 2.3% by FY2010. Erik Sherman, "Under Mark Hurd, HP Innovation Nose-Dived," CBS News, December 20, 2010, https://www.cbsnews.com/news/under-mark-hurd-hp-innovation-nose-dived/.
5. "Apple was at 4%": Apple spent approximately $489 million on R&D in FY2004 against revenues of approximately $8.3 billion. By FY2005, with the iPod driving rapid revenue growth, the percentage had moderated to approximately 3-4%. Phil's 4% figure reflects Apple's R&D intensity at the time of the Hurd conversation, contrasted against the PC-group-specific figures for HP and Acer. Apple Inc., Form 10-K, FY2004 and FY2005, available via SEC EDGAR at https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=AAPL&type=10-K&dateb=&owner=include&count=40.
6. "substitute the word 'bullshit' earnings": Charlie Munger's assessment appears in Poor Charlie's Almanack: The Wit and Wisdom of Charles T. Munger (Donning Company Publishers, 2005). He elaborated at Berkshire Hathaway's 2017 annual meeting, calling the metric "absolutely insane" because depreciation is a real future cost that EBITDA treats as if it does not exist. Erin Fuchs, "Charlie Munger Explains Why He Called a Popular Earnings Measure 'Horror Squared,'" Yahoo Finance, May 8, 2017, https://finance.yahoo.com/news/charlie-munger-explains-called-popular-earnings-measure-horror-squared-164228068.html.
7. "destroy a company's innovation capability": IBM CEO Samuel Palmisano stated publicly in September 2010 that Hurd had "cut out all the research and development" at HP, forcing the company into expensive acquisitions to compensate for lost internal capability. CBS News reported separately that HP's patent activity "sank like a rock" from 2005 onward as R&D fell not just as a percentage of revenue but in absolute dollars. Erik Sherman, "Under Mark Hurd, HP Innovation Nose-Dived," CBS News, December 20, 2010, https://www.cbsnews.com/news/under-mark-hurd-hp-innovation-nose-dived/. See also Erik Sherman, "IBM CEO Slams HP, and That's a Really Bad Sign," CBS News, September 15, 2010, https://www.cbsnews.com/news/ibm-ceo-slams-hp-and-thats-a-really-bad-sign/.
8. "retail shelf space HP had spent thirty years building": The Apple iPod+HP partnership was announced by Carly Fiorina and Steve Jobs at the Consumer Electronics Show on January 8, 2004. HP's distribution network included over 100,000 retail outlets worldwide, giving Apple access to Walmart, RadioShack, and Office Depot where Apple had no consumer electronics presence. "HP to Stop Selling iPods," Macworld, July 28, 2005, https://www.macworld.com/article/176488/hpipod-4.html. See also "iPod+HP," Wikipedia, https://en.wikipedia.org/wiki/IPod%2BHP.
9. "Apple was launching the iPod": The original iPod launched in October 2001. At the time of the HP deal, Apple's retail store network had only opened in May 2001 and carried limited consumer electronics reach. The HP channel gave Apple's product presence in mass-market retail during the iPod's critical growth phase. Luke Dormehl, "Today in Apple History: HP's iPod Comes Out of the Blue," Cult of Mac, January 8, 2026, https://www.cultofmac.com/apple-history/hewlett-packard-ipod.
10. "HP's complete financial history going back to the 1940s": Hewlett-Packard was founded in 1939 in a Palo Alto garage by Bill Hewlett and Dave Packard. The company's founding notebooks and historical records are preserved at the Computer History Museum in Mountain View, California. Computer History Museum, https://computerhistory.org.
11. "Acer today: approximately 0.9% of revenue in R&D": Acer has maintained R&D investment consistently below 1% of revenue for decades, a reflection of its cost-efficient, volume-driven PC market positioning. For current figures, see Acer Inc., Investor Relations, https://www.acer-group.com/ag/en/INV/investor.html.