Price metrix moneyball for advisors nov 2012
Published on: Mar 4, 2016
Transcripts - Price metrix moneyball for advisors nov 2012
PRICEMETRIX INSIGHTS VOLUME 7, OCTOBER 2012
Use data and analytics to drive
2 PriceMetrix Insights – Moneyball for Advisors
The application of statistics to professional sport was made famous by the book Moneyball: The Art of Winning
an Unfair Game by Michael Lewis. The book was subsequently made into a movie, with Brad Pitt playing the role
of Oakland A’s general manager Billy Beane. A central question of both the book and the film was: how could a
small-market, small-budget baseball team like the Oakland A’s compete against teams from larger markets that
had the resources to pay top dollar for superstar players?
The answer, in brief, was to challenge conventional wisdom and the intuition of baseball experts about how to
predict future success and build a winning baseball team. The Oakland A’s conducted a data and analytics-driven
exercise. Using statistical analysis, the A’s identified player statistics that were most predictive of winning
baseball games (and a winning team was one that baseball fans paid to see). Some of the metrics that baseball
insiders valued – such as runs batted in and bases stolen – were naturally priced high in the market for baseball
talent. They were also mediocre predictors of games won. At the same time, other metrics – such as on-base
percentage and slugging percentage – were more predictive of winning games, yet were underpriced.
By targeting overlooked players who scored well on underpriced metrics, the Oakland A’s were able to build
a winning baseball team less expensively.
Does the Moneyball revolution in baseball have anything to teach retail wealth management? To be sure, there
are a number of high-level similarities between the two. Highly-ranked talent has proven particularly mobile,
moving from team to team (or firm to firm) as incentives dictate. As a consequence, recruiting top talent has
become an increasingly expensive proposition: in the case of the retail wealth management industry, it is not
uncommon for high-performing advisors to be offered a multiple of two times their annual production to join
a new firm. An entire cadre of recruiters that focus exclusively on financial advisors exists, and news of
high-profile advisors moving from one firm to another has become a common occurrence. A recent Reuters report
noted that even in an era of tight budgets, “brokerages are often willing to pay top dollar for advisers with
significant client assets.”1
Like professional baseball, however, the payoff for the hiring firm (or signing team) remains uncertain. With
incentives based upon some combination of past performance and future performance, the payoff is difficult to
foresee; and with prices getting higher and higher, selecting the right talent is more important than ever.
A brokerage industry veteran recently observed that the economics of recruiting advisors “has gotten to the point
where it’s very difficult to continue to make money.” But still, “you have to do it from a competitive standpoint.”
Some hires, however, turn out to be based on a “bad assumption” about future production, which is obviously
detrimental to the business.2
The corollary for recruiting is to direct their recruiting resources and cash to advisors
with the greatest prospects for outperformance, and for advisors to concentrate their time, a scarce resource, on
those factors that drive outperformance..
Ashley Lau, “Hard Times at Brokerages Spell Good Times for Headhunters,” Reuters, October 16, 2012.
Tom Stabile, “Wirehouses Will Take Fight to Indies: Exec.” Fundfire, October 23, 2012.
Several PriceMetrix clients wanted to know if statistical analyses could be deployed, much as they are in baseball,
to help them ensure – rather than simply assume – that they are recruiting the right candidates and allocating
resources to those advisors with the greatest potential.
Research undertaken by the PriceMetrix Insights Lab confirmed that analytics can indeed be used to gain
an informed advantage in recruiting and development.
When it comes to recruiting financial advisors, a hire that is seen as a “success” has certain requisite qualities,
including the absence of compliance issues and a healthy margin. In this paper, our focus is on production (and
growth in production over time) as a hallmark of a successful hire. The answers we provide can assist
recruiters in identifying those candidates most likely to add to the future growth of the firm. It can also assist
managers and advisors by highlighting those attributes most predictive of future growth in an advisor’s book.
These insights can help advisors and managers take action now to best achieve breakaway growth.
This report is made possible by PriceMetrix aggregated retail brokerage data representing 6 million investors,
500 million transactions, 1.6 million fee-based accounts, 7 million transactional accounts, and over $3.5 trillion in
investment assets. PriceMetrix combines its patented process for collecting and classifying data with proprietary
measures of revenue, assets, and households to create the most insightful and granular retail wealth
management database available today.
All results are reported as of December, 2011. PriceMetrix found no significant difference between the U.S.
and Canadian markets, so all results presented in this paper are for the combined North American market.
4 PriceMetrix Insights – Moneyball for Advisors
DATA AND METHODOLOGY
The model we developed was designed to predict future revenue as a function of advisor and book metrics from
a benchmark time period. To do this, potential predictors were drawn from end-of-year 2006 figures for advisors
in the PriceMetrix database with 5 to 20 years experience (this range was selected as being representative of the
majority of advisors changing firms). The outcome being predicted was the twelve-month total revenue for 2011,
five years beyond our baseline metrics.3
Obviously, forecasted production is not a function of a single factor. Many different factors may potentially have
an impact – including advisor production in a baseline period, demographic characteristics and information on
advisor client mix, among others. We examined dozens of attributes of a book and characteristics of advisors,
in search of those that most powerfully predicted future production – many more than appear in this paper.
Not all proved to be predictive, and so do not feature in the final model.4
The end result is a robust, highly
predictive model, comprised of a small number of predictors, that offers clear guidance on what factors determine
future production. Managers can use this predictive model to help select preferred candidates in the recruitment
process, or to coach their existing advisors to perform better. Advisors can use it to groom their books in order to
ensure they have the highest likelihood of growing and succeeding.
PRODUCTION AS A MEANS TO PREDICT FUTURE PRODUCTION
Not surprisingly, advisor production in a baseline period is highly predictive of production in five years’ time. This
is what we would expect, since the norm when recruiting advisors is to offer incentives that are based in part on
current levels of production.
At the same time, not all types of production – transactional, trailer and fee – are created equal. It turns out that
baseline-period fee revenue is more predictive of future revenue than either transactional or trailer revenue.
This is again, what we would expect given that fee revenue is considered more resilient and stable than either
transactional or trailer revenue.
After controlling for the effects of other predictors, the link between benchmark-period production (2006) and
production five years into the future (2011) works in the following way. First, assume an advisor with average
production levels in each of transactional, trailer and fee business in 2006. Next, we systematically vary the
amount of fee-based revenue. Now, assume two advisors with average scores on all other metrics but with
different levels of fee revenue in 2006. One is below average with $100,000 in fee revenue in 2006; the other is an
outperformer with $700,000 in fee revenue in 2006. The predicted indexed performance5
of the first advisor is
-12; for the second advisor, it is 62 (see Figure 1).
Revenue data are client account revenue only; revenue from pro accounts and net interest margin were excluded.
An insightful, actionable forecasting model should also identify how much of future production can be uniquely attributed to each of the individual predictors. All of these considerations
point in the direction of using multivariate statistical tools to examine the unique effects of predictors while controlling for – meaning parsing out, statistically – the effects of
Indexed performance refers to scaling the model-adjusted mean total revenue for 2011 to be equal to zero. Scores above zero therefore represent above-average performance; scores
below zero represent below-average performance.
What this points to is the importance of examining the different sources of revenue in an advisor’s book when
evaluating prospects. Those with more production coming from fee accounts (as opposed to transactional and
trailer revenue) are expected to be bigger future producers.
We can conclude that, while current production can indeed help predict future production, not all production
is created equal.
IN RETAIL WEALTH MANAGEMENT, SIZE DOES MATTER
One continuing challenge faced by financial advisors is to acquire (and then retain) new clients – preferably with
significant assets. Most advisors believe that the greater the number of high-asset households in a book, the
better. Our predictive model confirms this belief.
After controlling for other predictors, we discovered that for every high-asset household (defined as those with
$250,000 or more in investable assets) in an advisor’s book, predicted future annual revenue is expected to
increase by $1,650. To illustrate this, consider again two advisors with average scores on all other metrics but who
have different numbers of high-asset households in their respective books – one with 50 (below average), and the
other with 200 (an outperformer). The advisor with 50 high-asset households is expected to have indexed
performance of -5 in 2011; the advisor with 200 high-asset households is expected to have indexed performance
of 27 (see Figure 2).
Figure 1 - Fee Revenue (2006) and Production (2011)
300 500 700100
Fee Revenue ($000s) 2006
6 PriceMetrix Insights – Moneyball for Advisors
Our analysis also found that the number of small households (defined as those with less than $250,000 in assets)
had a significant effect on future production – a negative one. Advisors actually pay a penalty in terms of
decreased future revenue for the small households they keep in their books. For every small household in an
advisor’s book, predicted future revenue is expected to decrease by $270 per small household, per year.6
the average advisor with 173 small accounts in his or her book, this works out to a total small household penalty
of more than $46,000 annually. Controlling for other predictors, an average advisor with 100 small households
is expected to have indexed performance of 3 in 2011; an advisor with 400 small households is expected to have
indexed performance of -8 (see Figure 3).
These results highlight the importance of client mix in an advisor’s book. Large households contribute to increased
future production. On the other hand, the effect of keeping small households in one’s book is malignant, as
previous PriceMetrix research on small households has shown.7
Advisors actually pay a small-household penalty.
Figure 2 - High Asset Households (2006) and Production (2011)
High-Asset Households (more than $250,000 in assets) 2006
Figure 3 – Small Households (2006) and Production (2011)
Small Households (less than $250,000 in assets) 2006
This reflects production that otherwise would be expected to have been realized if that small household had not been in the book.
PriceMetrix, “Small Household Metrics,” Insight, June 2010.
DEPTH OF RELATIONSHIPS MATTERS TOO
Marketing research experts have long known that the depth of client relationships and client “share of wallet”
are strongly linked to increased client retention and profitability. The results of our analysis align with these
expectations: advisors with deeper client relationships (measured using a number of metrics) are predicted to
have higher future production.
One good measure of the depth of client relationships is the number of retirement (IRA/401(k)/RRSP) accounts
in an advisor’s book. Previous PriceMetrix research has highlighted the contribution that retirement accounts can
make to an advisor’s book8
. Our analysis has also revealed that the number of retirement accounts in a book is a
predictor of future production. Indeed, after controlling for other predictors, each retirement account increases
expected annual future production by $510. To draw this out, consider once more the case of two advisors with
average scores on all other metrics but who have different numbers of retirement accounts in their books.
An advisor with 50 retirement accounts (which is below average) is expected to have indexed performance of
-5 in 2011. An advisor with 350 retirement accounts (an outperformer on this metric) is expected to have indexed
performance of 15 (see Figure 4).
Another simple measure of the depth of an advisor’s client relationships is average accounts per household
(simply the ratio of accounts to households in an advisor’s book). Higher accounts per household imply deeper
client relationships and a higher “share of investable assets” on average in a book. Accounts per household also
prove to be predictive of future production. Controlling for other predictors, an advisor with a below-average
accounts per household score of 1.5 is expected to have indexed performance of -4 in 2011. An advisor with
especially deep client relationships and an accounts per household score of 3.5 is expected to have indexed
performance of 9 (see Figure 5).
Figure 4 – Retirement Accounts (2006) and Production (2011)
Number of Retirement Accounts 2006
PriceMetrix, “IRA & 401(k) Accounts: Good for your clients, good for you,” Flash of Insight, February 2012.
8 PriceMetrix Insights – Moneyball for Advisors
Another key finding from our analysis is the conditional relationship between an advisor’s median household
size and the depth of client relationships. The common wisdom that it is good to have a book with a high median
household size is actually contingent: high median household size is a positive predictor of future production only
when accounts per household is also high. When accounts per household is low – implying rather shallow client
relationships – high median household size is actually a negative predictor of future production (see Figure 6).
Figure 5 – Accounts per Household (2006) and Production (2011)
Accounts per Household 2006
1.5 2.0 2.5 3.0 3.5
Figure 6 – Median Household Assets by Accounts per Household (2006)
and Production (2011)
Median Household Assets ($000s)/ Accounts per Household 2006
This analysis serves to highlight some of the important but often overlooked predictors of future revenue –
namely those that capture the depth of an advisor’s client relationships. It also highlights the risk inherent in
a book – not in terms of the risk associated with securities held, but in terms of the durability (or lack thereof)
of an advisor’s relationships with her or his clients. It is the risk of customer churn.
A book with few retirement accounts and low accounts per household, all other things being equal, actually
contains a higher element of risk than one with a large number of retirement accounts and a high number of
accounts per household. It also provides direction to advisors: households with significant assets are attractive,
but the medium to long-term benefit (in terms of production) is only realized if one can capture a high proportion
of that household’s share of investable assets.
THE BENEFIT OF INEXPERIENCE
Common wisdom might suggest that a more experienced advisor would be more productive, and therefore a more
attractive prospect for a hiring firm. However, our analysis showed that this is not the case. In the first instance,
the data showed no clear relationship between experience and future production. Before controlling for any other
predictors, experience had a negligible effect on predicted future production. Below-average producers, average
producers and outperformers are found in roughly equal numbers at all levels of experience.
More interesting, though, is what the data revealed when the relationship between experience and future
production was tested while controlling for the effects of other predictors. When other factors are included in the
model, the effect of experience on future production is negative. For every year of experience, future production
is predicted to decrease by $12,700. To illustrate, consider two advisors who are equal in all other respects – for
example, average producers in a benchmark period, average number of large and small households, average levels
of depth of client relationships, etc. In this scenario, the advisor with less experience has greater future potential
for increased production.
Specifically, an advisor with 5 years of experience is expected to have indexed performance of 11 in 2011.
An advisor with 20 years of experience is expected to have indexed performance of -13 (see Figure 7).
10 PriceMetrix Insights – Moneyball for Advisors
Simply put, an advisor who becomes a top producer after five years is a more attractive prospect than an
advisor who reaches the same level of production after 20: the advisor with five years experience is on a higher
Figure 7 – Years of Experience (2006) and Production (2011)
Years of Experience 2006
CONCLUSION: THE ART OF WINNING IN RETAIL WEALTH MANAGEMENT
What should recruiters pay for, managers coach to, and advisors do to maximize outperformance?
Our research has identified the characteristics of a book that most accurately predict future production.
• Maximum fee revenue
• A high number of households with greater than $250,000 in assets
• A low number of households with less than $250,000 in assets, and a willingness to exit or transfer them
• Deep client relationships, as measured by a large number of retirement accounts and multiple account households
• Experience – all other things being equal, less experience means greater potential in the future
Recruiters can ask prospects probing, specific questions to reduce the uncertainty around expected future
performance of new hires. Firms can improve their return on recruiting spend, by identifying ‘undervalued assets’.
Managers can use this research in the evaluation of their current advisors, in order to improve the trajectory of
future overall performance:
• Start by screening all the existing advisors in your sales force.
• Identify the future outperformers and hug them.
• Coach to areas of weakness with likely underperformers.
Advisors themselves can understand the factors that have the biggest influence on their future performance.
While advisors obviously can’t turn back the clock on length of service, other factors – such as fee-based revenue,
client mix and depth of client relationships – are within their control.
As an advisor, what should you do to ensure you are best positioned for future growth?
• Acquire ONLY households with greater than $250,000 in assets
• Consider exiting or transferring to a different service channel, those households with less than $250,000 in assets
• Establish a plan to cross sell your households, starting with retirement accounts
• Focus your new product sales energies on fee and managed business
Undoubtedly, there will be some advisors with an orientation to transactional books, small households and low
wallet share penetration that are able to outperform. There are always exceptions. However, the facts show that
the odds are, alas, against them. Those that play the odds will more assuredly outperform.
If you’d like to understand your firm’s “capacity” for future growth, please contact PriceMetrix. Our Advisory Team
can help you understand how your firm’s current advisor base will grow, how that growth will compare to your
market peers, and how the growth breaks down, advisor by advisor.
For assistance with understanding the opportunities your firm, branch, and/or book of business represents, or to
provide your feedback on this issue of Insights, please contact Patrick Kennedy, Vice President, Product and Client
Services at +1 (416) 955-1728 or firstname.lastname@example.org.
The analysis delivered in this edition of insights is
made possible by our aggregated market data and is
the result of a collaborative effort by Patrick Kennedy,
Vice President, Product and Client Services and
Tim Gravelle, Principal Scientist & Director, Insights Lab
This document and all of the components and content thereof (the “Research”) are propri-
etary to PriceMetrix Inc. and subject to copyright and other intellectual property protections.
All external or commercial citations of the Research are prohibited without our express writ-
ten permission. Contact Amrita Mathur, Director, Marketing at PriceMetrix at 416-955-0514 or
send an email to email@example.com to obtain our approval for any desired citations.
PriceMetrix reserves any and all rights to the Research including but not limited to the right to
deny any and all uses of the Research. The Research is provided only as information to readers.
By making the Research available, PriceMetrix is not engaged in rendering any commercial
consulting advice or services to the reader. All information and content of the Research is
provided without warranty of any kind and PriceMetrix assumes no liability for any reliance in
making decisions thereon.