Over the years I have tested almost every leadership assessment available. I have found them surprisingly unhelpful. Some are too esoteric - a page of percentages, or a diagram that renders my leadership as a shape, usually a slightly distorted diamond. Others are too simple, and only ever hand back my own view of myself. Neither kind tells me much I do not already know.

That is a problem. The point of an assessment is to show you where you are succeeding and failing in ways that are not obvious to you. Some of what we most need to learn about our own leadership is exactly the part we cannot see. An assessment that mirrors your self-image, or buries its finding inside a number, does not help you lead better.

I recall three distinct moments when an assessment changed my behavior. The first was in a class three decades ago. As part of a new student team, we gave each other feedback, and I was told that I kept saying "I" when I should have been saying "We". It was subtle, and it was useful. I felt like an outsider on that team, and without noticing it I was letting that show - the "I" was landing as a sign that I did not see myself as one of them.

The second came about a decade later, from a 360. My team told me they wanted me to ask them, more often, what I could do to improve my leadership and our team dynamics. I had thought that was mine to work out on my own.

The third was last year, and it came from a tool we built ourselves - the Decision Lab Leadership Pattern tool. It named something no assessment had ever put to me so plainly: a pattern in how I under-use my own authority. Seeing it set out that clearly was enough to make me rethink decisions I would otherwise have made on instinct. That is 3 crucial insights out of maybe 25 different assessments.

I am both fascinated by leadership assessments and determined to build ones that are insightful. Assessments are blunt tools when faced with complex leadership behaviors and no simple normative calibration - I mean the same behavior by a leader (for example giving strong direction) in two different contexts can be perfect in one and disastrous in another. However, diagnosing leadership capacities (yes I am using diagnosing deliberately) is crucial to be able to determine what to strengthen and what to adapt. An assessment gives you evidence. A diagnosis reads that evidence in context, across several sources, to work out what pattern is operating.

So, over the past few years I have designed and tested a number of new instruments and approaches to answer a range of interesting questions, and each draws on a different kind of evidence - from what other people report, to the reasoning a leader gives, to what they actually do when the situation is live.

Questions such as:

1. How do others experience you?

Our 360 gathers feedback on a leader from many people around them - their manager, their peers, and the people who report to them - so the picture comes from several directions at once. Much of its value turns on the questions it asks. A generic set of questions surfaces generic answers. So we curate the questions ourselves, matching them to our own model of leadership, Insight 4D, and to the leadership seniority you aspire to.

Any 360 has a limit worth understanding though. Each person sees only part of a leader. Your boss sees the strategy and the results. The people who report to you live with how you lead them day to day. A peer knows what you are like to work beside. None of these views is wrong. All are calibrated against what these people expect of a leader - which may or may not turn out to be helpful.

2. Why do you make the choices you do?

The Insight Decision Labs get at the reasoning under a choice. Instead of asking you to rate yourself - "I am good at giving feedback" - they present a hard situation and ask what you would do, and why. What surprises leaders is usually the assumption that made their choice feel obvious. You watch your own thinking as it happens, and you see the patterns that build across different scenarios. This is the tool I described above that has made me rethink a number of leadership decisions in recent months.

3. Are you actually changing over time?

That is what the Insight Leadership Behaviors Assessment follows. It tracks a range of complex leadership behaviors - collaborating across teams, say, or holding a crucial conversation without damaging the relationship - and asks two separate questions about each: how often you do it, and how well. The two move independently, and can provide interesting insights. You can be capable at something and still do it rarely. You can do it constantly and have it land badly. A lower frequency can even mean a higher standard, when a leader has handed work to the team and steps in less. In this assessment you rate each behavior before one of our intensive leadership programs and again after, which shows where you see yourself changing.

And in its strongest form we set the Insight Leadership Behaviors Assessment against a short team feedback survey on the same behaviors. You might believe you delegate often, and delegate in a way that grows your people. But does your team agree?

4. What do you do when the situation is live and the right decision isn't clear?

The Insight Leadership Flight Simulator is built for that. It is a practice tool and an evaluation tool at once. A leader works through a realistic scenario that unfolds in real time and has to make real moves as it does. It runs about an hour or two. The simulator captures the choices you actually made and your rationale for them, which is something a self-report or a colleague's later memory can never show. Afterwards we debrief those choices and their impacts, and the assumptions underneath them. This tool is one I am very excited about - and think it will be a strong accelerator in building capacity to lead through complex adaptive challenges.

Those three moments set the bar. An assessment earns its place when it shows a leader something they could not see on their own, clearly enough to act on. We are still building these instruments, and that is the test each of them has to pass.

Nonetheless, even the best of these instruments only gathers evidence. What it means still depends on the sophistication of the faculty reading it, holding the whole picture together and working out with the leader which patterns matter most.