5 ms·

Ask HN: Tips for a Big Data consultancy

9 points by rudimk 14y ago

I'm thinking of starting up a small consultancy, using Big Data for governments in South Asia and the Middle East. Mostly have coders in-house, who can analyze government datasets, and then report findings/analyses. As far as I can see, the single-biggest roadblock would be the ability to tap into governments - usually, there's an old-boy network, which we'd have to be a part of. What would you have to say? Any tips on making this idea better, and so on. Thanks a ton!

shankar1221989 14y ago
Who is your first customer? Build your product around him.
- rudimk 14y ago
  That's the thing - there's no "product" as such. Or, to be more precise, the product is purely consultative, the kind of work that, say, a think-tank does. The use case is, that a governmental agency needs help with analyzing a large dataset. We write the code, we do the heavy-lifting, and then present them with the final analysis.
  - grumps 14y ago
    He's not that far off. You'll have to build your relationship off your first client. The relationship will be your product. What's your consulting experience?
    - rudimk 14y ago
      Okay. Yes, that sounds about right. About 2 years of consulting gigs, with IBM and Cvent(event-management software) - but I've also worked with a think-tank, so I have a bit of insight into policy analyses and so on.
      - grumps 14y ago
        Are you going for a developer type role?
        rudimk 14y ago
        From a personal standpoint, no - I'll be working towards building relationships with mandarins. But at first, I would be writing code myself, as well as writing up reports and stuff. Eventually, I'm looking at a scenario where we have coders, and we have people interfacing with governments, reporting on our findings.
        grumps 14y ago
        You're just going to have to land your first gig, then expand your way from there. You'll have to knock them dead, and slowly bring in more people. Sidenote - Sounds like you're in the DC area, and if you're going for government work, I hope you have a lot of solid connections because 2-3 years of experience isn't always enough to land much more that a contract workerbee position.
        rudimk 14y ago
        Okay. That sounds a little daunting, but hey, not going back now. Actually..I'm not in DC. I'm based out of New Delhi right now. My take is, that the lack of such a consultancy in South Asia and the Middle East would mean, that my thing could really work here. Is that being naive, or am I actually on to something? Thanks a lot for the offer, though!
jasbanwait 14y ago
Agree with everyone, you need to start building your network with government contacts - start off with attending networking events and other related events to start talking to people about it.
- rudimk 14y ago
  Okay - that sounds like a good idea. Thanks!
apapli 14y ago
To me, "big data" looks to become one of the most over-used words in tech for 2013. So if you're planning to consult in this space you're probably onto a winner :) In addition to the good advice already offered in this thread, consider hitting up some of the prevalent consultancies in the region who already have government contracts. Subcontract directly or offer them a "big data analysis as a service" - some proposition that allows them to "tick the box" with their customers without having to invest too much. They'll already be on the "approved government list of suppliers", which means you'll be able to go from idea to revenue much faster. You can then move to independently sell to government after you've got a reference-able case study or two under your belt.
- rudimk 14y ago
  Thanks :) The interesting thing is, that atleast in India and Pakistan, the private sector has never been "allowed" to work on large-scale government contracts. For instance, all think-tanks in New Delhi and Islamabad are government projects, where people are still stuck in that Cold War mentality: two superpowers, and India's stuck in the middle. Unfortunately, while the world has changed, these views haven't. This was just an example of the prevailing situation here. So that's the problem - what I'm attempting to do is actually a lot more difficult, simply because the kind of partnership you'd see in the West between companies and organizations like Stratfor and Lockheed, and Western governments, simply doesn't exist here.
merinid 14y ago
You seem very casual about this. This is a line of work which is at the forefront of industry right now. People like http://www.k2intelligence.com/ http://www.k2intelligence.com/ and Palantir are going to beat you to it, but that's besides the point. Do you have knowledge in distributed systems? Do you even know what trends to look for? Have you played with Petabytes, or even Terabytes of data before?
- rudimk 14y ago
  Actually, as it so goes - I do. I majored in history and computer science. I've worked with IBM, on a contract with Amazon, harnessing EMR for analyzing customer sales, in order to figure out prevailing market patterns. Apart from that, I've freelanced as a policy analyst for think tanks in the Middle East, which in turn, collaborate with the State Department. My apologies for the apparent casualness - but my view is that even if I - or anyone else for that matter, doesn't know much, should it matter? It musn't serve as a barrier to doing something new. As long as one is ready to learn, though. That's the clincher.

9 points by rudimk 14y ago

shankar1221989 14y ago

Who is your first customer? Build your product around him.

rudimk 14y ago
That's the thing - there's no "product" as such. Or, to be more precise, the product is purely consultative, the kind of work that, say, a think-tank does. The use case is, that a governmental agency needs help with analyzing a large dataset. We write the code, we do the heavy-lifting, and then present them with the final analysis.
- grumps 14y ago
  He's not that far off. You'll have to build your relationship off your first client. The relationship will be your product. What's your consulting experience?
  - rudimk 14y ago
    Okay. Yes, that sounds about right. About 2 years of consulting gigs, with IBM and Cvent(event-management software) - but I've also worked with a think-tank, so I have a bit of insight into policy analyses and so on.
    - grumps 14y ago
      Are you going for a developer type role?
      - rudimk 14y ago
        From a personal standpoint, no - I'll be working towards building relationships with mandarins. But at first, I would be writing code myself, as well as writing up reports and stuff. Eventually, I'm looking at a scenario where we have coders, and we have people interfacing with governments, reporting on our findings.
        grumps 14y ago
        You're just going to have to land your first gig, then expand your way from there. You'll have to knock them dead, and slowly bring in more people. Sidenote - Sounds like you're in the DC area, and if you're going for government work, I hope you have a lot of solid connections because 2-3 years of experience isn't always enough to land much more that a contract workerbee position.
        rudimk 14y ago
        Okay. That sounds a little daunting, but hey, not going back now. Actually..I'm not in DC. I'm based out of New Delhi right now. My take is, that the lack of such a consultancy in South Asia and the Middle East would mean, that my thing could really work here. Is that being naive, or am I actually on to something? Thanks a lot for the offer, though!

jasbanwait 14y ago

Agree with everyone, you need to start building your network with government contacts - start off with attending networking events and other related events to start talking to people about it.

rudimk 14y ago
Okay - that sounds like a good idea. Thanks!

apapli 14y ago

To me, "big data" looks to become one of the most over-used words in tech for 2013. So if you're planning to consult in this space you're probably onto a winner :) In addition to the good advice already offered in this thread, consider hitting up some of the prevalent consultancies in the region who already have government contracts. Subcontract directly or offer them a "big data analysis as a service" - some proposition that allows them to "tick the box" with their customers without having to invest too much. They'll already be on the "approved government list of suppliers", which means you'll be able to go from idea to revenue much faster. You can then move to independently sell to government after you've got a reference-able case study or two under your belt.

rudimk 14y ago
Thanks :) The interesting thing is, that atleast in India and Pakistan, the private sector has never been "allowed" to work on large-scale government contracts. For instance, all think-tanks in New Delhi and Islamabad are government projects, where people are still stuck in that Cold War mentality: two superpowers, and India's stuck in the middle. Unfortunately, while the world has changed, these views haven't. This was just an example of the prevailing situation here. So that's the problem - what I'm attempting to do is actually a lot more difficult, simply because the kind of partnership you'd see in the West between companies and organizations like Stratfor and Lockheed, and Western governments, simply doesn't exist here.

merinid 14y ago

You seem very casual about this. This is a line of work which is at the forefront of industry right now. People like http://www.k2intelligence.com/ http://www.k2intelligence.com/ and Palantir are going to beat you to it, but that's besides the point. Do you have knowledge in distributed systems? Do you even know what trends to look for? Have you played with Petabytes, or even Terabytes of data before?

rudimk 14y ago
Actually, as it so goes - I do. I majored in history and computer science. I've worked with IBM, on a contract with Amazon, harnessing EMR for analyzing customer sales, in order to figure out prevailing market patterns. Apart from that, I've freelanced as a policy analyst for think tanks in the Middle East, which in turn, collaborate with the State Department. My apologies for the apparent casualness - but my view is that even if I - or anyone else for that matter, doesn't know much, should it matter? It musn't serve as a barrier to doing something new. As long as one is ready to learn, though. That's the clincher.