As a company and its industry mature, things change naturally. Often, growth becomes less important than profitability. Different metrics may emerge into the spotlight. Companies gain a clearer sense of which businesses fit and which don’t. They grow more likely to return capital to shareholders. In some cases, executives get more of their pay in cash and less in equity.

All of those transitions apply to Broadcom, an $8 billion company that’s second in the fabless semiconductor space behind Qualcomm. On Dec. 9, which the company dubbed Analyst Day, it outlined in detail the elements of its ongoing reset as a stable-growth company focused on profitability and shareholder return, following a long period of high growth — at times breakneck growth — that began with its 1991 launch. Impressed analysts offered praise and positive forecasts.

The enhanced transparency on Broadcom’s strategic initiatives came at the end of a year in which it decided to shut down its $750 million baseband chip business, aimed primarily at the lower end of the cellular phone market. It was a forthright move: R&D costs for that unit were out of whack with financial performance, and what had been an energetic and promising growth initiative was abandoned. The company retains strong positions with its infrastructure business (mainly Internet switches) and its broadband business (chips for home electronics, higher-end phones, routers, the Internet of Things and more).

Eric Brandt: value seeker

Finance chief Eric Brandt revealed on Investor Day the metrics Broadcom is now using to evaluate the performance of its 22 sub-product lines, in line with the company’s reshaped identity. Instead of the compound annual growth rate of revenue, it’s looking at gross profit CAGR. In place of weighing the sum of gross profit over a three-year period, its attention is on where each business ranks among the 22 in gross margin percentage. And rather than track each unit’s rank in research and development costs as a percentage of the company’s total R&D expense, Broadcom is now rating “R&D efficiency,” comparing each one’s R&D spending to its operating margin. The results of the ongoing evaluation of such metrics heavily inform decisions on capital outlays.

Brandt, who’s not known for granting a lot of interviews, made an exception for CFO last week. He told us about what the company is doing, where it’s going and his role in getting there. An edited transcript of the discussion follows.

Your focus is less on growth now, but the company is still growing, with the top line up 3.7% in 2013 and 5.3% for the first nine months of 2014. Do you still have growth goals?

I don’t think any executive is ever satisfied with the top-line growth rate, except for unique years when you’re up a lot. But in context, Gartner projects that the semiconductor industry should grow about 4% to 4.5% a year. As a company, our view is we want to be 20% to 25% faster than the industry. It’s hard to pick an absolute number in an industry like ours that’s cyclical, but if the industry grows 4.5%, I want to grow 6%.

Were we thrilled with our growth rate last year? No. But we substantially outperformed our peers in the broadband space, growing in the vicinity of 10 points faster, and in switching, which was about 7 points faster.

Where we had an issue was at the low end of the smart-phone market. As that market grew, the companies with higher market share, like Qualcomm and MediaTek, sold platforms, or a combination of chips for things like radio frequency and power management in addition to connectivity. So that’s where the growth was, and since we were not participating significantly in baseband, we missed a fair bit of that growth. In the handset space we now sell connectivity as an individual component rather than a complete platform.

You shut down the baseband business in late July. What impact has that had on the company as a whole?

It was an annual $700 million investment for a business that wasn’t generating enough revenue to cover that. So our operating margins have improved. Instead of 20 to 22%, we’re targeting more like 23 to 28%, in both cases on a non-GAAP basis, or excluding stock-based compensation.

Why are you focusing on different metrics now?

Between 1985 and 2003, the semiconductor industry was growing between 10 and 30% a year. Now it’s more like 4 or 4.5%. So it’s gone from a new market to an evolving market to a stabilizing market. Understanding the economics of that is key to running the business.

It’s not that we are focusing on new metrics “now.” This has been a multi-year journey for the company. The exit from cellular baseband, which had the potential to add almost 50% to the company’s revenues, was the Rubicon for formally acknowledging the change.

As a company in an industry that’s growing 4.5%, if you aspire to beat that by a certain amount you have to have a profitable growth focus. That’s what drives you down the path of a different set of metrics, moving from growth at any cost to growth that drives economic value. The prior metrics rewarded size and revenue growth; the new ones are skewed much more to value and return, which feeds earnings per share.

How do you use the metrics?

We stack rank the businesses on each metric. That provides a portfolio grid that helps you understand the choices you’re making, particularly as you’re allocating R&D dollars.

I try to do things very visually so it’s easy for everybody to understand. If you’re in the top third for any metric, you’re color coded green. The middle third is yellow, and the bottom third is red. Now, you might expect that different businesses perform better or worse on different metrics. But in the top third, most of the businesses are green on most of the metrics. In the middle third, most are yellow with a little bit of green or red. At the bottom, most of the metrics are red. That allows you to be more comfortable about prioritization, so that you’re not overweighting any particular metric.

Can you explain the change in the R&D metric?

To arrive at R&D efficiency, we take the operating margin and subtract the R&D percentage. It recognizes that there are some lower gross-margin businesses with less R&D that might be very attractive businesses, and some higher-margin businesses that use higher R&D and are also attractive businesses. The issue with cellular baseband was that it was low gross margin with high R&D.

When businesses with high gross margin and low R&D want incremental R&D dollars, they should get it.

Either way, Broadcom is spending a lot of money on R&D — 30% of its revenue in 2013. That sounds like what a startup would do!

I came from the pharmaceutical industry, which most people think is the most R&D-intensive business in the world. But the semiconductor industry is far more so. Putting aside the strictly analog companies and the memory companies, just looking at those like us that do digital mixed-signal work, most are in the 20% range. If you wash out some of the GAAP charges, excluding the baseband business we’re at an R&D rate of 24 to 25%. That reflects the complexity of the solutions we typically sell and the gross margin we produce.

Among some of the other changes you’ve made, you have more performance-based compensation in the executive pay mix, and the portion that’s equity-based has been going down for years. What’s that about?

When you’re a young, entrepreneurial company you give out a lot of equity principally because you are balance sheet-constrained. You don’t have cash, so you give out equity. As a larger company, you’re not balance sheet-constrained, and a more cash-based compensation system has better economics to the overall business. We used to grant close to 14% of revenue in stock, and nobody does that. Now we’re closer to 4%, which is more consistent with the industry and our peers.

As to the performance-based compensation, four of the executive vice presidents had principally been getting time-based stock, with only a small piece, 25%, tied to performance, in the form of revenue growth and operating margin. Now we’ve taken them to 50% performance, while the senior vice presidents have been bumped up to roughly a third performance and the vice presidents to 20%. They’re all tied to EPS growth and total shareholder return.

Why the switch to EPS and TSR as the basis for performance pay?

What drives value, particularly for a mature, asset-light business such as ourselves [fabless semiconductor companies design and sell chips but outsource their fabrication, so they don’t own manufacturing facilities], is really EPS growth, which is what drives TSR. The r-squared, or the correlation between earnings growth and stock price, is 80% for the S&P 500 [i.e., 80% of stock-price growth is attributable to earnings growth]. With a database of 500 companies anomalies can be washed out, but r-squared can be applied to a single company too, and ours is 60%. That indicates it’s an important metric for TSR. We are therefore transitioning executive compensation toward a series of metrics that correlate to stock performance.

So, what are your current priorities?

This year, the first third of the year was spent evaluating the baseband business and trying to figure out whether to stay in it, and then peeling the company apart in a way to explain the economics of the decision to shut it down. Then we had to exit all the people necessary in order to strip the company down to its new shape, which was followed by evaluating what the right metrics for the company should be, and then finally communicating it, which resulted in Analyst Day.

The priority now is getting the company on its new footing as quickly as possible. Periods of change create uncertainty, and uncertainty can damage a business, whether it’s because you lose key people, or customers get uncomfortable, or things slip with so many things changing at once. We have to cement across the company the management processes and thinking that the executive team has developed and begin executing against the strategy.

I think once we get past that there will be new things, new vectors we’ll want to look at. But our plate is pretty full in delivering on the pretty significant things we laid out for Analyst Day.