How easy is it to predict the success of CEOs and FDs?
The importance of CEOs and FDs to the success (or failure) of private equity backed companies is relatively self-evident. But to what extent can we glean useful clues from data regarding incumbent or candidate executives before either doing additional assessment work or making a decision? To explore that question, Catalysis has been updating its databases and analysing the relationship between data on CEO/FD experience and leadership styles on the one hand, and information we have on individuals’ subsequent performance in role.
Does experience predict performance?
We have detailed data on the career backgrounds of 163 CEOs and 80 FDs. These are self-estimates (with only occasional corrections by us when figures look mistaken) of the years an individual has had:
Across their total career
Working in this industry and/or these kinds of clients
Working close to these kind of offerings
Benefitting from systematic development
Exposed to good processes
As a manager
As a senior team member
As an equity holder in their employer
In this company
In the current management team
In private equity backed companies
Exposed to fast growth and change
We then cross-reference these with ratings of effectiveness in role using the following classification:
Unusually effective in playing their role (Shown as ‘Very high’ in the graphs)
More effective than most (High)
About as effective as others (Medium)
Less effective than most (Low)
Unusually ineffective (Very low)
Since the number of Very highs and Very lows for both CEOs and FDs is quite small (which increases the chance of distortion), we have combined those with Highs and Lows respectively to create decently sized groups (for CEOs 70 Highs, 58 Mediums and 35 Lows; for FDs 25, 26, 29 respectively).
To decide whether a difference between higher performing managers and their lower-performing counterparts is significant or not depends not only on the size of the difference but also whether the sequence of scores across high to medium to low managers is transitive. i.e. if the scores form a progression, which suggests that the dimension being considered might be relevant to their performance as opposed to just random variation.
The results of this analysis might seem counter-intuitive: there is little sign of higher or lower performing CEOs and FDs having significant differences in their high-level experience. High performing CEOs have had a year less experience, claim less industry experience, lower levels of exposure to development and good processes, and reckon they have had slightly less senior team experience. High performing FDs are broadly similar, although they have had slightly longer careers and slightly greater SMT experience.
Those unexciting results could suggest three types of explanation:
a. The data just isn’t solid enough for one reason or another (e.g. different understandings of the terminology, sample sizes still too small).
b. The results are correct – which could mean that age is a factor: for CEOs youthful energy might outweigh years of additional experience; perhaps FDs’ roles depend more on accumulated experience.
c. The results reflect lower performing CEOs showing greater willingness to exaggerate their experience to seem more impressive.
There is only one result which does suggest a major difference, strong transitivity and objectivity, as well as an intuitive explanation: the proportion of CEOs and FDs who are graduates. As the table below shows, performance seems strongly related to educational status.
We still need to be careful drawing implications from this (see a fuller discussion of the issue here https://www.catalysis-advisory.com/blog/qualifications-matter-27ywf) because there are plenty of high-performing CEOs without degrees and plenty of low performing CEOs and FDs who have degrees. Nonetheless, the correlation is striking.
Can we predict performance from personality dimensions?
The leadership profile Catalysis uses, Saville Assessment’s Wave, which has the advantage of offering plenty of detailed information on executives, specifically 108 dimensions (called ‘facets’ by Saville) which can be grouped into hierarchies. We have that data – and performance information – on 205 CEOs and 95 FDs. Crunching the data to look for significant differences and transitivity, as we did for experience data, we find as follows:
CEOs
Highly rated CEOs are more likely (than low rated ones) to score themselves highly on:
Analysing information; Numerical data; Probing questions
Develops strategies; Clear vision for future
Motivating people; Encouraging
Multi-tasking; Likes working under pressure; Finds ways to improve things
Sense of self-worth
By contrast, lower rated CEOs are more likely (than high rated ones) to score themselves highly on:
Recovers from setbacks; Takes optimistic view
Punctual; Cautious decisions; Honours commitments; Maintains confidentiality
Some of these clusters make intuitive sense. CEOs who are analytical, able to think about the future, skilled at dealing with pressure, inclined to engage others in their plans - and liable to self-confidence – all seem like features likely to promote performance in a private equity environment.
Likewise, perhaps the lower performers are too focused on compliance and too cautious to drive value. But their ability to recover from setbacks and demonstrate optimism feel like they should be positive predictors, not warning signs! In any case, the differences are generally not major and the explanatory power (R-squared for those know their statistics) of a model based on these items is worthwhile but nothing to write home about.
FDs
The equivalent data for FDs seems less intuitive. Highly rated FDs are more likely to score themselves highly on:
Explains things; Giving presentations
Establishes rapport quickly; Team oriented
Accepts change; Calm before events
Focuses on results
Sense of self-worth
Studies underlying principles
Meanwhile, lower rated FDs are more likely to score themselves highly on:
Definite views; Gets involved in arguments
Motivating people; Understands why others do things; Empathetic; Copes with upset people
Responsibility for big decisions; Self confident
Trusts intuition
Learning new things
Some of the positives where higher rated FDs seem relevant (change oriented; team oriented; sense of self-worth) – but not necessarily more compelling than those where lower rated FDs seem themselves as stronger (e.g. learning new things; various people related areas). Curiously, however, the explanatory power of a model based on this is stronger than for CEOs.
Improving the odds?
Would having this information on an individual potentially improve the odds of selecting a stronger rather than weaker CEO or FD? The answer is yes, apparently, at least to some extent. Using the personality dimensions to build a model, we can see how predicted performance would translate into actual performance for the executives involved here:
Our data suggests that none of the CEOs in our highest category of predicted performance went on to under-perform (in fact almost all were above average), whereas half of those in the bottom two categories did.
For FDs the picture is even starker: all of the top predicted people ended up performing above average in reality; the large majority of those in the bottom two categories under-performed.
Putting the pieces above together we can say that backing or hiring a CEO or FD with a degree and certain personality dimensions (especially at the higher and lower extremes of our model) is likely to improve the overall odds of subsequent performance. That is useful to know. However, the causal relationship is far from perfect and there is a chance that at least part of that relationship is caused by random variations.
So, what really matters?
In reality, of course, no-one would make an important decision based mainly on the high-level information presented above. Instead, reaching a decision about senior executives requires at least four other types of insight:
The absence of anything disturbing or toxic about someone. Catalysis works with our friends from Neotas [hyperlink] to check this.
More detailed information about the individual – such as motivations, mental horsepower, grit, self-awareness, insight accumulated from past experiences, past performance – which requires proper interviewing and referencing.
Good understanding of the business and organisational context in which the individual will need to perform. That requires proper specification for new roles [hyperlink to blog on specification]
Understanding of the team with whom the individual will work to generate results – how effective are they individually and collectively?
Conclusion
Using high quality data and tools can usefully help shape perceptions of likely probabilities of success at executive level. However, there are no silver bullets when it comes to making complex judgement calls about executives in growth companies. Instead, that activity requires patient piecing together of insights from multiple sources and synthesising them into a robust decision.
