Serving as a corporate board member, particularly for a public company, seems like an intimidating job. Good — it should seem intimidating, in a way that drives us to be better fiduciaries. But the role of corporate director is also a fairly simple one, if we remember what our essential job is — to hire and monitor the top executive team, starting with the CEO. The board that stays focused on this knows its role in the company, and is able to effectively oversee performance, strategy and culture… while still staying out of the CEO’s hair.

But like most big, simple job descriptions, the board’s task of assuring good leaders brings lots of complicated tactical issues. Just how do we measure our leadership (and for what)? What tools and approaches work best for CEO succession planning? And how do we best structure our company leadership?

Let’s start with how a board assesses the performance of its CEO. Board should take the lead on this by discussing the type of leadership the company needs. Schedule this review at least annually. The velocity of change is now so fast that many companies hold a multi-day, offsite retreat to shape a forward view of strategy in the competitive landscape. One obvious element of this is to look at current and future leadership, mapping it against your expected future dynamics.

Regarding when and how the board should think about replacing the CEO, start by looking at external data for benchmarks. The average U.S. CEO tenure is now at 9.7 years, according to Forbes. But that assumes a “steady state” company, rare in our era of accelerating change. Look at your structural factors. The lifespan of a company according to EquityZen is 11 years before it goes public or is acquired. This helps clarify the question — is the CEO that you have the right CEO for the future?

Assuring a CEO succession plan that aims toward the future can be challenging, especially for smaller companies lacking a deep talent bench. Data show that two-thirds of CEOs developed from within are successful, versus just one third of external CEOs. But small cap firms without a strong bench of internal candidates will need the board to extend the annual talent review discussed above, to think about if/when they may have to look for an outside hire. The current roster may just not be ready for the next stage of the company’s growth. Not all CEOs and team members can scale from a $100-500 million company to a far more complex two billion dollar one. There will likely need to be an external leadership search. I’ve found the best way to manage this review at the board level is to make it part of the compensation or governance committee’s annual talent assessment.

Speaking of talent assessment, given the stakes your board is dealing with, don’t just go with your gut instincts. My boards often use both outside and internal CEO candidate assessments, including psychological screening. These are often handled by executive search firms. Large-cap boards may have a deeper process orientation and formalized CEO succession review. These big companies typically seek an in-depth assessment of the internal candidates to benchmark them against external candidates. Potential leaders should be mapped against where the company needs to take its strategy over the coming three to five years.

For example, a company whose business model and strategy face little significant change, seeking just to execute the core business model through incremental growth and profitability, is likely best served by an internal candidate. Many companies, though, have business models radically impacted by coming change (Sheraton-Hilton lodging model versus Airbnb, or Carey Limousine impacted by Uber). These boards may need to look outside for someone with deeper tech or innovation skills, or fresh business model expertise to help the firm cross the competitive chasm.

The board’s leadership planning shouldn’t focus just on individual talents, but on how the role itself is structured. In the U.S., combining the jobs of chief executive and board chair has long been the de facto model. There has been some erosion in this leadership norm, though. Currently, just half of S&P 500 companies split the two roles, compared with three-quarters of British FTSE 100 companies (the FTSE has a “comply or explain” policy on dividing the jobs). In America, if the CEO is not ultimately named board chair, it’s viewed negatively. For example, in the case of a successful, iconic CEO such as Nike’s Mark Parker, it would be a seen as demeaning not to be awarded the chairman role.

In CEO succession, the point when a CEO retires and a new CEO arrives is a convenient window for splitting the CEO and chair roles. Naming a retired CEO to become chairman of the board is often a sticky situation though. The new CEO often feels second guessed by the former chief. The chairman/former CEO may have a very hard time letting go of operational responsibilities, and your board may have inadvertently set up a dynamic with high risk and many negatives. I would advise boards to think long and hard before they put a retired CEO into the chairman role.

This dynamic grows even more complicated when a retiring CEO is also the company founder. A founder who retires as CEO to take on an “emeritus” board chair slot is a formula for very high risk. There are very few examples where this has worked successfully — and many more where it has hurt the performance of the company and significantly distracted the new CEO.

When boards agree to this structure, I fear they’re often not facing the really hard decisions. It’s always awkward and difficult for a board to transition a founder. They feel ownership of their companies, and have many deep relationships. It is extraordinarily difficult for a founder who is emeritus chairman to stick to the “boardroom” parameters, not overstepping and inserting him/herself into operations. In weighing a founder emeritus chairmanship, boards must ask themselves if they truly are the one in 100 (or even one in 1000 case) that will work. Consider carefully the risks of destabilizing the new CEO for the shareholders. A better approach is for the board to shape a “special advisor” role for the founder, assigning special, targeted projects. One example might be to define a one year role to specifically assist in some predefined areas in the transition to the new CEO.