Get a Management System — Now

This is a continuation of the series that began with Startups Need Management, Too

tps report2.jpgAll the warnings about the forthcoming next nuclear winter for startups drive home the need for a management system. Tough times call for tough measures, but that doesn’t mean CEOs should get carte blanche for discretionary cuts — a sure way to undermine morale.

Among other things, having a management process in place sets the stage for making cuts when they need to be made. Why? Because even without resorting to old-school ranking-and-rating, in a good management system, performances (and non-performances) are visible company-wide.

I’m still stunned at how many companies operate without a management process of some kind. Not just a plan . . . but a regularly recurring process, with weekly monitoring, that ensures critical company milestones are met — and leads to consequences if it’s not.

So if you haven’t got one, now is a helluva good time to get started on a planning/management process. If your company has one in place, dramatic events may call for a mid-cycle re-plan.

The argument for doing quarterly planning is its horizon — it’s hard to see much beyond 13 weeks. Who can really say they know what conditions will be like come January 1, 2009? Much less for a 3- or 5-year plan.

(Which is not to say your company doesn’t need long-term goals. VCs usually only want to see 3- to 5-year plans as a reality check (or sanity check). They know better than anyone that even the startups that succeed rarely do so by adhering to the plan they funded. But it’s a good test for founders, to articulate whether they have their sites set on a profitable $15M business two years out . . . or are going for global domination of mobile advertising by Year 5.)

Back to quarterly planning — and in particular the elements of Quarterly Objectives and Key Results that I brought up in the last post. Although the process needs to evolve within any company, the basic premises are three:

1. Top-management expectations come first. Could be just a one-sheet with simple but clear expectations for, say, sales, shipments, user registrations, feature releases — whatever the major performance metrics for the quarter are. Whether or not there’s a board of directors asking for it, it’s the CEO’s job.

2. Every exempt employee prepares a set of Objectives & Key Results. If the company is big enough to have departments, employees work with their managers to develop them — four to six Objectives with four to six Key Results (more on these below).

3. After each quarter, Objectives are reviewed and self-graded, and next quarter’s are proposed, in a peer setting. Key here is that each employee acknowledges what happened last quarter — and what they’ll do next quarter — in the presence of their peers.

Now, some are you are sitting there saying, “Sounds like big-company bullshit.” To which I say, “Bullshit.” (Touché!) Do you really think everyone working hard, with a good attitude is a way to ensure milestones are met? To ensure that roles, much less departments, are coordinated? No. Not only is there too much to keep in any one’s head, but things change. All the time. If you have your eye on the goal — pin it to the wall in front of you — and you monitor weekly, big surprises are far less likely.

“Inch by inch is a cinch — yard by yard is hard”

Corny, but it makes a point: monitor progress every week.

I won’t suggest that even this ‘simple’ process I’m recommending isn’t pain-free. There are usually several iterations, especially when multiple departments (engineering, marketing, sales) dependencies are getting sorted out (“I can’t launch Week 44 unless Tom agrees to freeze code by Week 40”)

But at the end of the exercise, every employee — the CEO is not exempt — has a roadmap of the most important things for them to focus on over the next 13 weeks.

I can’t emphasize how effective this process (or one like it) is in not only getting things done, but in maintaining employee morale. There’s no hiding, no slacking — everyone’s work is visible. No more “What the hell does Chris do?” Everyone’s Objectives are on the server (or intranet, or Basecamp, or whatever you’re using). And every week at your staff meeting (what — you still don’t believe in meetings?), all the Objectives (sorted by week) are checked off . . . again, in front of peers. (Peer performance is a great motivator — no one wants to show up for Monday morning’s meeting and get marked ‘Not Done’ in front of others.)

What the Docs Look Like

QO pic.jpgThough few words, drafting of Objectives and Key Results takes some time to get the hang of. In general, I would describe each employee’s four- to six-page document as:

‘A guide which, if 75% to 85% achieved, would signify successful execution for the quarter.’

Why only 75% to 85%? Because everyone should have stretch goals . . . and therefore should not be expected to hit 100% of them.

So how is each Objective expressed? It should be broad, categorical, and have no dates. Examples might be ‘RELEASE v1.5,’ or ‘SECURE MANUFACTURING PARTNERSHIP,’ or ‘GROW SALES AT LEAST 10% Y/Y,’ or ‘COMMENCE PATENT PROCESS,’ or ‘EXPAND DEVELOPMENT TEAM.’ They’re allowed to sound vague — they’re the Big Efforts, and each individual should have at most six. Any more, and people are being set up to fail. Put each one at the top of a separate page.

The steps — specific milestones — to achieving each Objective — are quantified by Key Results, which are measurable, and include dates (in our work-week format). Again, try to have four to six of them. So your CTO’s Objective ‘RELEASE v1.5’ might have the following Key Results: ‘FREEZE CODE W40,’ ‘COMPLETE QA W43,’ ‘COMPLETE USER TESTING W47,’ ‘FINAL REVISION W49,’ ‘BLOG RELEASE NOTES W50, and ‘UPLOAD v1.5 W51’ (although a Christmas release may be unduly harsh on your development team).

The idea is to give focus your very busy people on the few items — Objectives — that really matter. An while Objectives are mostly vague, Key Results use specific action verbs — ‘COMPLETE,’ ‘HIRE,’ ‘ANNOUNCE,’ not ‘Continue,’ Investigate,’ ‘Study,’ etc. — so that there’s no ambiguity as to what ‘Done’ means.

There’s much more than can be spelled out here. And many ways to skin this cat. (Apologies to cat lovers.) Finally, while this is an example of a process to help keep the company’s execution on track, there’s another dimension (which I’ll address in a subsequent post): people management. They’re not the same.

Especially in these dark days, this is intended to get you thinking about taking real steps to get your team to pull together and sharpen execution — even if there are only two of you. Exercises like this force you to articulate your thinking, and you’ll be surprised how often your partners, whom you talk to every day, are surprised.

If you’d like to receive a more detailed description and examples of Quarterly Objectives and Key Results, email me at rcapece at technosailor dot com.

Startups Need Management, Too

Grove High Output Management (FINAL).jpgProspectors joining the dot-com gold rush in the ’90s were mainly coming from large organizations seeking to capture some of the new wealth. But along with the promises of stock options and casual dress came another bonus — no bureaucracy. No meetings! No Microsoft Exchange! And no more onerous management systems.

Disciplined startups recognized that management systems were important — for setting and hitting milestones, and for giving employees adequate frameworks to perform and feel good about themselves and their company. But more often there was extreme swinging of the pendulum that led to free-wheeling, hair-on-fire mismanagement, wildly missed targets . . . and lots of disgruntled campers.

Driven by the mantra of ‘first to market wins,’ dot-com startups were hiring way too fast, pushing employees way too hard (I remember one of our VCs saying he expected all employees to keep a sleeping bag at the office), and eschewing management systems entirely. In the end, a lot of folks went back to their big companies, happy for a return to structure, sanity, and their 401k . . . even with a pay cut.

The fact is, every organization needs a management system. It just needs to be appropriate for the company’s stage.

Much has been written about motivating employees. The admirable folks at web-app developer 37 Signals (makers of Basecamp and other popular utilities) espouse a four-day week and flexible hours, and for the most part, I ascribe to their philosophy. But I also have seen lots of companies provide increasing perqs to increasingly dissatisfied lots of employees.

What it gets down to is that providing the right amount of structure, goal-setting, and feedback — and communicating clearly — does more for esteem and spirit than all the free food and Friday afternoon keggers ever will.

After a six-month stint consulting for a startup to help it transition from a service to a product business, I joined the company full time as COO. It had about 20 employees. The CEO had good transparency — employees got monthly updates on progress and direction — but individuals had tasks, rather than goals, and no way of seeing how their roles fit into the bigger picture.

This is one of the key tenets of management: people want to know how their contribution fits in — how their efforts (along with their counterparts’) ‘roll up’ in support of the company’s overarching goal.

One of my charges was to institute a planning and management system. And it’s honestly one of the most satisfying aspects of management — especially when you see people responding . . . when they come by at the end of the day to tell you ‘I really appreciate what you’re doing here,’ or ‘we really needed this.’

Startups’ management systems (and culture) are usually brought by the founders from their antecedents. Most of my management experience came from a manufacturing startup. My co-founder came from chip-leader Intel Corp. — along with a fairly high percentage of our initial hires — we essentially followed Intel’s management disciplines, policies, and procedures. In hindsight, it was one of the smart things we did. Intel was an extremely well run company. (I subsequently became a huge fanboy of then-CEO Andy Grove and his management books.)

I adopted Intel’s Quarterly Objectives and Key Results methodology, and have used it (with a few of my own variations) ever since.

To be sure, shipping a manufactured product requires a different discipline than the ‘just get it out there and revise it later’ strategy of web apps. But certain management principles are universal, and over the course of working both in hardware and software companies, I came to understand which ones they were.

One of the principles that always took a while to convince people of was the work-week calendar. Lots of them are available, but in particular, many manufacturing companies split the year up into four 13-week periods . . . usually with ‘4-4-5‘ months. (Mainly to ensure a ‘linear shipping’ schedule, making it easier to hit quarterly shipment targets and to compare quarter-to-quarter performance.) But it’s something I found to work well in hardware, software, and Internet businesses.

Why? Because over time — it usually takes three quarters or so — the numbered weeks start to stick, and people in the company realize that it’s a lot easier to reckon you have 11 weeks till year end (we’re in Week 41 right now) than calculating the number of days between October 8 and December 31. More importantly, employees and managers get into a recurring 13-week rhythm, which has certain psychological advantages.

And why is it so important to think in terms of quarterly performance? Whether you’ve just got plans to make money, hit revenue and profit targets, and grow — or serious ambitions to become a public company (IPOs will return someday!) — ‘making the numbers’ each quarter is a discipline that should begin early.

Next post: The Basics of the Quarterly Objectives process.