The last Word, The President , God and Bill Gates


I use three social media networks – Facebook, Twitter and Linkedin. I blog too – so maybe that is a fourth channel. 

Facebook is primarily for personal interests – keeping in touch with family in India and elsewhere, keeping track of what is happening in the world of dog shows and keeping in touch with friends from school. I have a handful of work peeps there, but that is just coincidence. 

Linkedin is my address book for the most part – and for the most part, I use it to get a curated list of good reading material.

Twitter falls in between – I have no intentional personal/professional divide there. But if I am forced to categorize, it is skewed more towards work than personal. 

For the most part – being active in social media has only been helpful to me both on work front and personal front. I used to think that they have very little in common though. Facebook keeps a conversation in one place, which is hard to do in twitter despite hashtags etc. Linkedin has been the least engaging for me – but that is because I don’t get into groups all that much despite signing up for a few. And twitter has been awesome for me despite all its flaws – except for a few people who just incessantly post famous(?) quotes, forcing me to unfollow them.

But as time progresses – I have realized that I was rather wrong in thinking they have not much in common. I think there are three common characteristics in social media irrespective of the specific channel .

1. “The last word is mine and mine alone” syndrome

There is an undeniable urge people have to always be the one who gets the last word in. It gets downright silly – and at least in my circles it is the same whichever way you segment (age, gender, education, country, or pretty much any thing you can use to divide a group into two). There is no denying that I have done it too.

In real life, more people try to make a point and walk away gracefully when they are heard. That is not “a thing” in social media. In social media –  a good majority stay and fight every nitty gritty endlessly and ensure that the whole conversation goes south and all relationships sour. And then they rinse and repeat. It does not matter what the topic is – if they don’t get the last word, they will keep at it with absolute tenacity. 

2. “I hate the current president” syndrome

It does not matter whether the president got elected democratically just like your guy did the last time – in social media, hate for the current president is perpetually a hot topic. While the tempo increases at election time (which is expected), it does not exactly go away after elections. Granted there is some election or other at any point .

If anyone had any doubts on whether the country is divided at a random point in time – just logon to FB and read for an hour. People – including some analysts that I think have the most balanced views on enterprise – have no sense of balance in their political beliefs. 

Now I have lists of such people that I just mute for a few months around election time .

3. “God and Bill Gates loves or hates you depending on what you share” syndrome

About a fifth of all posts on my feeds in facebook are from people who share something saying god will punish you if you don’t share in next five minutes, or from people who are convinced that Bill Gates is going to send them a big fat check for sharing something about his software or his foundation. In fact – I think there is way more of these folks than there are Nigerians promising to make me rich via email. If you don’t believe that a sucker is born every minute – check face book. You can thank me later 🙂

 

 

Large IT company business models and the cloud


Earnings season arrived and I am wondering how long big companies can keep their current structures and business models . Size is not always an advantage – it can actually destroy the value of a company .

Big companies try to attack this problem by firing a lot of people . But I have hardly known a company that has succeeded much that way – only very few pull that off .

The problem with firing enmasse is that you have to use broad policies to determine who needs to go . When you fire 1000s of people – not only do you end up firing people who could have done wonders , you lose other good people who will leave because they feel they can do better elsewhere . Firing people lower down in the hierarchy rarely does any company any good. Those are the peeps who do actual work – and they are not usually the ones who made the decisions that turned bad and led to management needing to fire people . Micro corrections throughout the year are a whole lot better in my opinion than mass firing . But at large size and matrix organizations – you cannot optimize . I know that first hand – I have tried and failed , repeatedly .

Large IT companies that have HW and SW have a problem – HW becomes a leading indicator for software , when investors look at the whole company . So if HW declines now , market assumes software will suck pretty soon too . The worst part is if you also have services in the mix . Services trail both HW and SW . So for such companies – if they don’t stop bleeding for HW, market will probably punish them for a long time in future .

The move to cloud is what seem to screw up the larger companies . Cloud is a low margin business – which a large company doesn’t really want to be in . If they don’t do it – they will become obsolete . If they do it – it takes a large amount of time for market to absorb the shock of lesser top and bottom lines . Damned if you do and damned if you don’t . Essentially cloud economics seem to work only for companies who don’t do anything else but cloud business .

When companies solely depend on EPS as a way to set goals for investors – they just lock themselves in jail and throw the key away . To beef up EPS, you can only do three things . Make more revenue , spend less cost and buy back shares . You can’t make more money in cloud – it’s commodity business compared to legacy revenue streams . Your marketing and data center and retraining costs will make sure you can’t spend less . So the only way they can survive is by buying back shares aggressively . That is money they can’t invest back in business. As soon as investors see it – they punish the stock . It’s kind of funny in a sad and weird way how this tactic seems to boomerang

So how exactly do they come out of this vicious cycle ?

I think the sustainable way is to sell off low margin businesses as stand alone entities and double down on what you know best . And if that doesn’t work – then split the company into smaller entities that don’t have to bear huge over heads because of checkers and double checkers . I don’t think any big IT company board is ready to do anything that drastic . They can try to innovate their way to glory – except it is very rare for innovations to scale to an extent that it compensates for the drag caused by deteriorating legacy businesses .

So I guess I will just keep wondering what will happen to all these goliaths . Is there a silver bullet somewhere ? A lot of people I care for work in those places and for their sake , I hope good things happen to those companies

MongoDB for the SAP Community


One of the most common questions I get from my friends from the SAP ecosystem is how MongoDB works with SAP. So I thought I will pen some initial thoughts on how I see the technologies working together. 

NOTE : Please don’t look at this as an official roadmap or anything – this is just my personal view. Neither MongoDB nor SAP has agreed to do anything of this sort . This is just Vijay thinking aloud – and I am sure someone will set me straight if I am not making sense.

First, what does MongoDB not do with SAP systems ?

This is easy – MongoDB is not a database to run your ERP / business suite or BW type systems. MongoDB also doesn’t do native predictive analytics.That is what Hana (or other RDBMS systems) are for. MongoDB has no competitive position against Hana. Also SAP Hana Cloud platform has been using mongodb for some time already for content management. And MongoDB is one of the databases for SAP’s Hybris platform.

Now that we got that out of the way, lets see what can MongoDB do to complement an SAP landscape

1. Visualization

SAP announced a technology collaboration for Lumira to work with MongoDB at SAPPHIRENOW last month. Lumira is SAP’s agile BI solution . MongoDB holds data in JSON (BSON internally ) . The existing connector uses a JDBC protocol for the integration – and more native options are planned for future.

What is the big deal ? the big deal is that MongoDB holds richer data than what you typically see in rows and columns of a table. 

MongoDB holds data that has a hierarchical or nested nature – and visualizing that rich data ( for example, not only what region that customer is from – but details of what is the past history of sales in that region for customer, what other customers have bought in similar categories etc) is an amazing value add for a business user.

On top of that – given the dynamic schema, you can keep enriching the data (like campaign information at real time from a web site) and visualize the results immediately . 

2. ETL/EAI

I could totally see in future SAP data services and Hana cloud integration for ETL/EAI use cases where MongoDB is the source or target. MongoDB is probably one of the most  popular database for cloud applications already – and that trend will only increase. But those apps need data to be fed into it from somewhere, and I think data from there might need to get pumped into other systems. So an ETL/EAI type tooling from SAP around MongoDB is something I can imagine being a sensible thing assuming those sources and targets could be SAP systems.

3. As a datahub for integrating SAP and non-SAP applications

Consider a business application SAP would build on Hana for retail . The sources of information that need to pump in data to that app will typically be disparate in every sense – POS data, ERP data, IS-Retail data, Web store data, Click streams and so on.

One way to do it is to pump all this data into Hana and write code there to filterthe dataset that the app needs. This will make sense for cases where the schema is set in stone and not subject to frequent change. But that is usually not the case for retail like data – schema needs as much flexibility as possible. So a second way to design that will be to use MongoDB as a data hub to consolidate data from all the sources, and then from there provide the app on Hana with just the right data it needs to do its thing. 

There is no certified connector for MongoDB with Hana today. But that probably is not such a big deal for a project based solution once the Hana SDK for AFL (app function library in C++) is available. MongoDB has a C++ driver that can be hooked on to AFL  and this should solve the data exchange problem from a technology point of view.

Another option – a little bit of code on XS engine (the lightweight web server/app server in Hana)  can make it work too – since XS understands JSON well. Thomas Jung probably is the best contact to help with this – but I don’t think you will need a lot of support. 

And for non-Hana customers of SAP, you probably need to code in ABAP from the app server layer to make mongodb integration work. I have not opened an ABAP editor in a while – so my memory is rusty. But I think its only from 7.x that ABAP has native JSON support. So you might want to do some hack to check if Kernel supports JSON. Horst Keller might have published some standard way to do that. Or if your ABAP skills are not top notch and you would prefer to use PERL or something outside the app server that MongoDB has a driver for . I guess you can create a service via ICF and take an RFC route. Probably not the most elegant way, but it should work too. Maybe Thorsten Franz can enlighten us on the possibilities.