Analytic(s) Aint for Martians

Spread the thoughts

And you do not always need to hire a ‘specialist’ to ‘do’ Analytics.

In fact Analytics (whatever value add that means for business as usual) is inseparable from ‘transactions’. It always was. Tool or no tool, Math or no Math, structured or unstructured, internal or external data, SMART managers ‘do’ Analytics in their minds (and hearts) all the time – 24 x 7. Give them some raw inputs from systems of record and a spreadsheet (or a pen and paper) and they will come to the right conclusions swiftly. They would have already gathered enough ‘offline’ intelligence from the market (or wherever) by then. An amount of information that you perhaps cannot store in your 100 TB data warehouse (if there exists one). Only condition is the manager should have spent a reasonable amount of time in the organization. With happiness and sense of purpose. They can discern, describe, visualize, predict and act, ALL at the same time such that the next deal with the next partner is profitable enough. They can see though the patterns that a ‘sophisticated’ Analytics software perhaps cannot. Be it a marketing manager (e.g. which promotion to discontinue effective tomorrow), credit manager (who is likely to default next month, who not to lend to), supply chain manager (what is the probability of out of stock situation in April, what will be the capacity overload situation in March), Demand Planner (e.g. which marketing manager is over enthusiastic and how much is the bias?) or finance manager (e.g. what will be the cost of capital next quarter and what can it mean to bottom line in scenario A, B and C) or Human resource manager (e.g. how much incentives to pay an unhappy overworked manager to make him/her happy this year). They make decisions all the time FOR the sake of ‘transactions’ they need to do. Any additional insights and inputs to such managers is a bonus. Smart that they are, they find the information they need. They scan the environment all the time. They are also cross functionally sensitive. They are in touch with market. They speak to people of all kinds all the time. A great sales manager is also a great human resource manager!

Then there are ‘dumb’ Managers who get their ‘subordinates’ to create ‘intelligent’ reports from ‘transaction (ERP)’ systems and then embellish it with all kinds of colors. A system that perhaps readily offers the kind of filtered ‘lists’ at the click of a button (worst case with some outer joins on couple of tables to query it out.. IT can help) but these managers won’t move their posteriors one inch to help themselves. Even with the most 'detailed’ ‘reports’, they would demand yet another ‘column’ to reassure themselves that they are not making the wrong decision (e.g. how much to replenish to that DC now). These are information hungry managers. They won’t take decisions swiftly and any additional ‘analytics’ to such a manager is a recipe for confusion and further delays on needed action.

But there is a much larger confusion peddled by makers and marketers of ‘Analytics’ software. I wonder how so many firms out there manage to sell their ‘Analytics’ software that does no more than some descriptive ‘computations’ that even a reasonably good high school student can do for no money. Some of these software are positioned as if it is some highly ‘advanced’ software not meant for ‘ordinary’ people of ‘ordinary’ firms with tiny little minds. This is serious marketing issue. You do not really need an analytics software of the kinds peddled to compute percentages, ratios, averages, variances and biases. You can compute these numbers with a B&W calculator or an excel spreadsheet that costs less than 50 dollars in licenses on a life time basis.

Secondly they market Analytics as a disjointed computational activity. That is not helping anyone. Definitely not for traditional manufacturing-centric firms. Practically of little help to Managers if the firm (CIO) is tempted to create a new department of Analytics with a bunch of Quants from MIT after being impressed by a sales pitch by a hyper aggressive salesman. The way Analytics is marketed now (esp. to traditional ERP install base) conveys such a message. It is positioned as a specialist diagnostic ‘activity’ separate from business as usual. I have a problem with such positioning. It does not help with short term decision making process by those(users) who are expected to make the decisions NOW.

Thirdly there always was a specialist software and specialist purpose trained analysts in the market for certain ‘Analytically intensive businesses” e.g. portfolio analysis, risk analysis, securities analysis, search analytics. machine learning and such. There is an entire industry out there as distinct from the manufacturing firms where Analytics needs to assume a more practical do-able meaning. A meaning in the context of WHO is generating, analyzing, consuming, and publishing the ‘new’ insights. If the generator, analyzer, consumer and publisher is the same person then any Analytics for such a manager must be integral to his/her day to day decision making process. There is no luxury of a separate new box and server! No one has got the time even if there is money to throw

Then there is research based on quintuples of data. That job is best left to statisticians and management scientists for macro-economic and medium to long terms conclusions. Industry and domain Analysts as they appoint themselves need to take a lead here E.g. how will the new product category fare in the next 5 years? That needs a different skill set and tuning of mind. But some smart managers who have their fingers on the pulse can discern that too. The insights from the Analysts is a reassurance.

The good news though is Enterprise Software leaders KNEW this problem need very well. As it applies to firms that buy, make and sell something. They knew that Analytics (or Business Intelligence as it is called in inward looking sense) cannot be a disjointed activity or a department of publishers and disseminators of palatable information for decision makers post facto. Analytics on the fly off real time data - your transaction data + more external data if you care to buy and store is what is needed.

I am no expert to comment on technological limitations here wrt computing speed and presentation of the information but at least we can hope some meaningful developments here from SAP. E.g. SAPs HANA (as in-memory database + development platform that can port all kinds of Analytics on demand) holds some promise. What I do know is EVERY manager is unique and his/her appetite for slicing, dicing, viewing and discerning insights is unique. The flexibility that is available with popular UIs like excel must be available here too– add / delete / pivot / aggregate/compute/graph / model / add-ins/ publish etc. Without having to fly experts from Hong Kong and Philadelphia for a new ‘variable’ or a statistic that the manager needs now. UI’s matter but that is a subject of human computer interface and interaction, a vast area of research in itself. That is needed for the 'dumb' managers. But yes this is some good news only for the smart managers who need to quickly run some graphs and numbers while on the call with a prospect who expects some price and delivery assurance right now. You do not have to tell him ‘I will get back to you shortly’. You say it NOW. Over the phone. What analytics does the manager need here? Do I have enough capacity to entertain this order at the price and date expected? Can I make a more attractive offer in the interest of conserving service promised to other customers? That needs data from a dozen data sources to run analytics on the fly.