I wrote this when I was with a large BPO firm, managing a team of "knowledge process outsourcing" (KPO) professionals performing business research, financial analysis, statistical analysis, and so on for global clients. The article also draws upon my experience in management consulting, where research and analysis is an essential part of the consulting process.
PDF copy:
http://f1.grp.yahoofs.com/v1/gCGrSZMft71gW_lyKRKva8nl1AMajMUPO-Id5IPIe4iDmDNK71smUaO7qRz04PQ1hgdK6M_v9iVCTAJOJEHlQw/Bluffing%20your%20way%20through%20research%2C%20May%202005.pdf
---------------------------
Bluffing your way through research
Partha Sarathy
So you decided to take up a career in business research, against the best advice of your friends and well wishers. Your life has been turned upside down, trying to grapple with vague client requirements, lack of data, insane deadlines and despotic bosses. Sounds familiar? Don’t worry, help is at hand. We have just completed our exhaustive study on what makes some researchers highly successful. Here are the nine techniques used by highly successful researchers that you can follow to impress the boss, overwhelm clients, get a raise and build a great reputation.
1. power of derived variables
When you are at your wit’s end trying to make sense of your data, derived variables can help you salvage the report. Say you have data on sales, market cap, number of branches and number of employees, but there is no discernible pattern or relationship. Then use derived variables like (employees/branch) or (market cap/employee) to get results that are intriguing and plausible but cannot be easily dismissed. Make it more exciting by comparing one derived variable versus another. If you are desperate, use compounded derived variables. Who can argue when you say that strategy 1 is superior to strategy 2 because the former yields higher (sales)*(P/E)/(employees/branch)?
2. Magic of matrices
A matrix has the power of a thousand words, so use a matrix if you are confused beyond words. Beginners usually draw simple ‘X versus Y’ matrices that anyone can understand (and challenge). Skilled practitioners know how to use abstract variables that infuse the matrix with authority, power and mystery – such as innovation quotient, revenue distance, and degree of verticalization. A good matrix is like modern art – you can admire and interpret it any way you want, but you can’t dismiss it. Here are some best practices on how to make use of matrices:
use 4x4 or 3x3 matrices instead of 2x2 matrices
assign derogatory names to one quadrant (dogs, laggards, mother-in-law)
manipulate the axis scales to ensure that no quadrant is empty
compare important variables with irrelevant/unheard of variables
At the hands of a great master, the matrix is elevated to truly sublime levels of truth, life, beauty and meaning. If only the almighty had an opportunity to consult with MCK or BCG before framing the ten commandments!
3. powerpoint for combat
PowerPoint is definitely the greatest human invention since the wheel and toothbrush. Used well, it becomes a weapon of mass destruction, creating shock and awe among the audience. The power of PowerPoint is its ability to present anything any way you want regardless of the content. You can create a 1-slide overview of a 1,000-page report, or (more frequently) create a powerful 20-slide presentation out of a 1-sentence idea. This is much like the separation of the presentation layer and application layer in software development – and unlike that other curious IT phrase, what you see is not what you get. Naturally, PowerPoint is very popular with VPs and top management. After extensive research, we have been able to identify the reasons why VPs love PowerPoint so much:
VPs inhabit very high altitudes in organization charts, and their bodies have evolved to use PowerPoint summaries to cope with oxygen deprivation
Given their rarified altitudes, VPs need to get a 30,000 feet view of what is happening in the company and the market
Viewed from 30,000 feet, the world looks like a PowerPoint slide
4. Box charts
Helps you represent and illuminate relationships between variables in a clear manner, regardless of how well you understand the alleged relationships. Classics include the 5-forces model of industry structure or the diamond model of competitive advantage of nations. Moral of the story: if you and I draw 5 boxes and throw in a few arrows the client would take us to court, but if a Harvard professor does that it’s called a ‘model’. The way you connect the boxes in your chart reflects what is really going on in your mind – for example:
A few boxes with minimal arrows (means “I have no idea what the heck is going on”)
Profusion of arrows connecting every box with every other box (“I am overwhelmed, I give up”)
Random arrows between some of the boxes (“I know the answer, but I want to preserve the mystery”)
Complex chain of boxes and arrows, like the structure of a particularly obnoxious chemical compound (“I am suffering from indigestion”)
5. Circular reference
This is a powerful technique that gives you the license to use concepts/phrases in a deliciously vague manner by defining them in terms of other equally vague variables, at least some of which are defined in terms of the original one. For example your report on reverse engineering may define it as a byproduct of something called process deconstruction, which is defined in a footnote on page 256 as a special category of reverse engineering. Circular reference in software will cause computers to hang, but management is happily immune to this malady. A variation of this technique is called non-existent reference, also called the Emperor’s New Clothes reference. You make wild claims in the first part of the report by saying “as we will prove in chapter 5, ---”, but chapter 5 only contains a few oblique references to the original concept, and later chapters breezily assert “as we demonstrated in chapter 5, ---“. Your chances of being able to get away with this trick depends upon:
length of your report: anything below 100 pages is risky
readers’ senility: Vice Presidents and above are safe
readers’ gullibility: MBAs are fine, but never try this with accountants and lawyers
6. Torture your data until it confesses
If you have lots of numbers that don’t seem to fit into any pattern, there are 2 things you should do so that your results agree with what you want to prove: (1) keep on trying different operating variables, and (2) use different analysis techniques. For example, if you compare net margins and get no clear results, try other variables like EBIT, EBITDA, operating margins, gross margins, contribution margins, and so on until something clicks. (Keep this in mind next time you see a report comparing Earnings Before Taxes and Depreciation, After Interest and Advertising Expenditure.) Also try analysing the data in various ways: averages, ratios, correlations, square roots, logarithms, exponentials, conditional variables, anything else you can think of. With these 2 techniques, you can prove anything using any data.
7. Research methodology
Great researchers do not waste time in pointless research, they are extremely results-oriented and client-centric. Their formula for success is the research methodology they use, which goes like this:
step 1: identify what the client wants to hear
step 2: decide what results you want to show
step 3: list down the hypotheses that could yield those results
step 4: identify what data you need for this
step 5: collect data
step 6: analyze data
step 7: write the report
Unfortunately, novice researchers spend too much time and effort on steps 5, 6 and 7, without thinking through the client-centric steps 1, 2 and 3. Gandhi must have had researchers in mind when he said: ‘have purpose, and means will follow’.
8. chinese numbers
When you estimate market potential for a new/unfamiliar product, always use at least 3 different methods that give wildly different results. This helps you to (1) choose the answer the client really wants to hear, (2) cover your posterior when years later the market size data is published (e.g. if you projected a market size between $1 billion and $100 billion, chances are you won’t be wrong). At least one of your methods should be based on the funnel principle, which starts like this:
country’s population is X million
number of households is Y million
spending on existing product A is $P per year
assumed proportion of households interested in new product B is Q%, and so on.
9. Know what the client does not know
When you present a report on an unfamiliar industry/topic to people who have been working in that area for 20 years, you have to focus on something your audience would never have heard about. Here are some examples that great researchers use:
case study of company X which faced a similar problem in Italy (if client is in Europe, use Mexico instead of Italy)
market size projections in China (the advantage of China is that you can talk of billions, trillions or quadrillions, and the audience won’t/can’t refute you)
(PEG ratio/Net margin) ratio for companies in that industry
impact of proposed EU regulation (if client is in Europe, mention NAFTA), and so on
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment