Business & Technology Nexus

Dave Stephens on technology and business trends

Performance Based Pay – Thoughts from a Silicon Valley guy

with 6 comments

What are the views of a Silicon Valley guy like me whose spent a decade in Oracle’s performance-driven culture? To steal a line from Fox News, I think you’ll find them “fair and balanced.” :)

Let me explain the context. In talking with friends and colleagues (like Michael Lamoreaux), it’s come to my attention that the vast majority of organizations aren’t as modern as my San Francisco Bay Area upbringing had me imagining. I thought corporate America and the business world overall had embraced the value of performance based pay. I thought businesses understood that by aligning a portion of a person’s compensation with corporate interests that the company grew more competitive over time.

But I really should have known better. After all, lack of a performance-driven culture is what led me to join Oracle in the first place…

Straight from university I joined a small engineering consultancy called SetPoint. It was a good company that had cornered the advanced multivariable process control market, at least for a time.

Over my 4 years there (1992-1996), SetPoint would face increasing pressure from an upstart called DMC. Both firms, it turned out, were eventually swallowed by AspenTech in the same M&A transaction.

SetPoint did have a bonus plan. And bonuses, to their credit, did vary based on your performance. Top performing engineers received, as I recall, around 6.3%. Lower performers, on the other hand, might get 5.8%. The difference was so slight that it was barely enough to cover a meal at a nice restaurant!

Now I will be the first to admit I was pretty full of myself back then. I was looking to prove what I could do, and so naturally I tried to take on as much work as I could. I eventually ran a team of consultants who did projects for GE, personally wrote custom software applications on those projects, modified core product functionality in cooperation with SetPoint’s products group, taught classes for the Education group, did Marketing work including working with legal on trademarks, etc. It was a classic fast-and-furious small company role.

But at some point fellow employees started stopping by my office and asking me “Dave, why are you working so hard. Don’t you know it doesn’t matter?” And that was tough to hear. It got me thinking about how insidious that attitude was inside the company. What happens when you are working in a firm full of people who are watching the clock and heading home right at 5pm? How vulnerable are you as a business? And who is having any real fun!?

I asked for a meeting with the CEO to better understand why the company wasn’t more performance-driven. To his credit, he accepted even though he didn’t know me. And at that meeting I learned more than I ever imagined. Here’s my 10-11 year old recollection:

Dave Stephens (annoying young engineer): “Doug, why isn’t going above and beyond rewarded more here?”

Doug White (CEO): “Dave, this is a company of engineers. Engineers value stability above all things. We don’t differentiate much on performance – that would make things unstable. We like it that way and we’re not changing things.” (but more polite and CEO-like)

I knew right then I was working for the wrong company.

At Oracle, performance was and is managed fairly well. Every 6 months teams pull together forced rankings and use them to determine bonus and stock incentive compensation. Now I’m sure there are many firms that do performance-based pay better than Oracle – I’m not declaring them the best by any means. Yet I’m equally sure many more firms do it “worse” & could learn from Oracle’s techniques.

But measuring performance can be extremely tricky. A week ago, I posted a teaser showing a full & empty parking lot at a manufacturing facility. The question was, if just as many widgets are output on Sunday with 50 people there, how come 400 people are required Monday-Friday.

It would be naive to think that the other 350 people aren’t needed. But for the 50 required to run the manufacturing line, measuring performance is straightforward. How many widgets? Oh, up 10% at the same quality level? Great job. And so you could conceivably incorporate production targets into their compensation plans. But for the other 350, you’re stuck.

Measuring engineering and product design, which I’ve had to do a long time now, is more like measuring the 350 workers at the plant – it’s difficult. Get too formulaic and you’ll incent the wrong behavior. Say you start measuring developer performance based on code line output. Developers will start optimizing for that, not quality. Measure on defects and you will destroy a collaborative and open working environment as people cover their behinds. Formulas and straight quantitative measures simply don’t work in my experience.

I think the best plans recognize that performance is subjective and don’t hide from the subjectivity and complexity in the measuring process. Performance factors should be broad, and of course must include enterprise performance as the number one factor. We were always told at Oracle to rank people based on “their contribution to the company.” And I thought that worked.

Another thing to recognize is performance-based compensation systems are never fair. You should still use them – but you should understand this point fully. Subjective measures are loaded with complexity to begin with, and then exacerbated as different managers value those measures in different ways. So you are never going to have a plan or execute a plan that’s 100% fair. The best you can do is try and approach fairness & correct mistakes as you make them over time. Managers owe it to their performance-driven teams to promote the goals of these programs while acknowledging their inherent fallibility.

By being honest & upfront and not getting overly specific, management can ensure performance based pay, just like procurement, is good medicine for your business.


Written by Dave Stephens

09/28/06 11:43 AM at 11:43 am

Posted in Historical, Opinion

6 Responses

Subscribe to comments with RSS.

  1. It’s really hard to do performance-driven reviews fairly. What if person A works less hard than B, but person A’s departure would be a blow to the group (for example, s/he wrote a chunk of mission-critical code that only s/he understands)? Can I really afford to antagonize A, and reward B? There’s only so much money in the “raise bucket,” and as a manager, I’m bonused on what I can deliver. If A gets disgruntled and leaves, my schedule is in trouble, so what do I do?

    Plus, if I reward only workaholics, what message is that sending? I become a sweat shop like some of the video game development houses we’ve been reading about. Reminds me of a course I took at MIT, where only one problem set in the whole class was given a “5”. If you got all the answers right, you got a “4”. But if you went nuts on the problem set, submitted all kinds of extra work, etc., you might get the coveted “5”. It was irritating and pointless and stupid (yes, I’m talking to you, Mike Hammer).

    Doing performance reviews has always reminded me of college admissions. You have to let some of the alumni’s C-student offspring attend, because otherwise you don’t get any big checks at fund-raising time. But at the same time, you can’t afford to admit too many slackers, or overall student quality goes down.

    The purest performance-based reviews are done by consulting firms. They are absolutely consumed by reviews (as they must be, because only the tiniest fraction of “associates” can ever be allowed to become partners). I’ve seen consultant performance review forms that are staggeringly complex. Peer reviews, manager reviews, partner reviews, customer feedback, self-analysis, critiques of the self-analysis, critiques of the critiques, and so on. This process can consume a significant fraction of the aggregate IQ of the firm.

    What seems to work best in my experience is when everyone owns a piece of the company, which usually ensures that all the oars are pulling together. Problem is, in a big company it’s awfully hard to divvy up the lousy few percent of equity that’s typically made available by owners/investors, so that it has any material impact for the recipients. And, the slacker problem can become an insoluble problem if an “owner” decides to stop pulling his/her weight.

    Sales compensation is another supposedly performance-based model, which rewards by revenue contribution. But, I have seen many examples of sales people (in the spend management vertical, too) being let go or forced out, because their commissions drive their income above that of jealous CxO’s or sales managers. The prospect of million-dollar years is always held out to sales people, but seldom does it occur more than once. Next year, the million dollar baby finds his territory split up (“you can’t possibly handle all these accounts by yourself”) or her compensation messed with (“for really big deals, we’re cutting the commission percentage way back”).

    Maybe Oracle got it right somehow, but I can’t say I’m overjoyed with any performance-based solution I’ve ever seen.

    Eric Strovink

    09/28/06 2:10 PM at 2:10 pm

  2. Here is why large companies can never get it right. Say a large company has a 6-tier hierarchy. Say a young D who is the best in his class-of joins this company as a leaf. Unless D->manager, D->manger->manager .. all the way up are also top performers, they won’t be able to get enough budget for their group and as a result young D gets screwed. In large companies like MSFT and ORCL, I would be surprised if a VP can match a face to the name of every individual in his/her organization! So, VP of D will not have a clue about D and so there is no way D would get recognized for his extraordinary contribution. I have seen several examples of people leaving the company because they don’t believe in their manager’s potential. Remember, young D will sooner than later graduate and learn the work politics/culture and move on. Ofcourse, history may repeat, but at a different company.


    09/28/06 6:36 PM at 6:36 pm

  3. 1st I’ll just agree and say I’ve never seen a performance-for-pay system that was perfect either. And no, Oracle didn’t somehow get it right.. I personally screwed up plenty of times – making mistakes in allocating and ranking. And often times the difficulty was just what the commenters say – that progressive management layers (up to a VP like me who did not know everyone by name in his organization, even though i tried) were progressively clueless about the people they were ranking. So it was far from perfect. What I’m left with though, is akin to the “democracy is terrible except when compared to every other form of goverment” type feeling. Performance rankings and performance for pay have all the problems they’re accused of, but I think they are still better than an hourly salary and a totally disconnected workforce. Straight equity in the company might be better, but of course that only works for us small guys :)

    Dave Stephens

    09/28/06 6:51 PM at 6:51 pm

  4. I agree with most of the comments here; compensation is always problematic in large corps; it is a lot easier for startups/smaller companies (Setpoint wasn’t too big and I think had 6-7 founders/owners so I’m surprised by Dave’s insights here;-)…

    To make peformance-for-pay system work well in a large org, it is critical that all top managers are “A” players and each has “ownership” in the company… In my experience, Bechtel – they are a $16+ Billion engineering-procurement-construction company – has done this quite well… Every SVP is a partner/owner in the privately held holding company (s-corp!) and virtually all senior managers are “A” players who

    With today’s technology, we can take the performance ranking to the next level – as we have done with our team members, a majority of whom are offshore. All our team members do force ranking of other team members, and grade them on a scale of 0-10, all via online surveys… Also, in addition, they distribute a hypothethetical bonus pool amongst the team… And, the client team members also provide similar feedback… Thus, we can do a lot of things that the Big-Consulting firms do, at a fraction of costs/bandwidths by doing it all via online surveys. Over past 12 months and 3 release cycles, we have followed this process, where final decisions are made by the team lead/manager and the cash incentives are 20-40% on top of above avg base salary… As a result, we have a high performance team incld the 15+ offshore employees…

    V C

    09/29/06 2:11 AM at 2:11 am

  5. Am not sure if this thread is still alive. But I desperately need some advise.

    I run a really small software company with 6 developers, 2 testers, 3 sales and support staff. In myr grand scheme of things, I hired, trained and created this core team for product development.

    Right now we are not profitable – hope to be so in the next few months. The demand for the skills that the team has is pretty high – essentially this means that they can leave when they want and finding a job will be no problem at all.

    Now I need a decent incentive based salary structure that motivates the small team without causing them to unduly compete against each other.



    01/2/07 10:42 PM at 10:42 pm

  6. for a small firm like yours, what i believe you want is twofold:
    a) provide options or straight-equity to drive “ownership” behavior
    b) recognize that you’re the Chief Marketing Officer as far as helping the team you’ve built see the bright future you’re small firm has..

    you may have already done both of these things but i cannot stress their importance enough.

    as for incentives, they should be pretty different based on roles. for sales staff, i trust you already have sales quotas established with significant bonus payouts or accelerators based on over-achievement. for support staff, i think some sort of periodic review of customer satisfaction with the support experience combined with new ideas that bring support costs down over time are the right 2 measures to offer incentive-based pay on.. for your testers, you should be careful not to simply measure defects found but instead incent based perhaps on a combination of defects found, developer satisfaction with the testers, and new ideas that reduce the testing burden.

    i’ve always found incentives around developers the toughest. i wrote some about this in my initial piece. my guess is over a reasonable period of time you’ll know who the best are – those should be candidates for nice bonuses and perhaps additional option consideration down the road.

    not sure if these comments will prove helpful. i wish you the best of luck in retaining your team & growing your business.


    Dave Stephens

    01/5/07 10:57 AM at 10:57 am

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: