April 18, 2013
Should Technology Ever Drive Business Strategy?
Examining new possibilities created by Big Data
There are many reasons why we generally don’t want technology to drive business strategy. Many companies have fallen prey to buying into buzzword-heavy or “shiny” technologies that either they don’t really need, are here today and gone tomorrow, or are simply not a good fit for their business requirements. So can it ever be ok, or even beneficial, to put the cart before the horse by letting technology advances drive a business strategy? For an example, let’s examine the evolution of data storage technology, including recent advances related to Big Data.
It used to be that, when building a data storage platform, we would setup a centralized relational database, perhaps establish a warm disaster recovery site using database replication or a mirrored SAN file system, and call it a day. Sure, we had a single point of failure, it required very expensive hardware, and it had limited scalability, but we just accepted that as normal because it was the way we were (and still are) used to building systems. Besides, there were large, profitable, influential, database vendors that advocated that approach. The consequences of failure due to not following their advice could be career-limiting.
Fortunately, there are always technology-driven companies who stand to benefit from challenging conventional wisdom. One relevant example that comes to mind is the adoption of Object Database technology by financial trading companies. When each millisecond of latency makes a difference to the bottom-line, there is tremendous incentive to push the envelope of data storage technology. Despite some important drawbacks, Object Databases often deliver massive performance gains compared to relational databases, and for financial trading companies, they provided a real competitive advantage.
So why didn’t more widespread adoption of Object Databases occur? Here are some likely reasons:
- The important business data was already in relational databases (negative inertia)
- The business could not clearly assess the risk of moving to other technical solutions (closed source products, lack of understanding)
- The technology was not widely implemented or tested in their industry (no industry champions)
- There were other alternatives to achieve similar benefits (namely, improved performance) that were generally “good enough”
Although Object Databases raised expectations of what a database should provide with respect to performance, they did not offer benefits that were compelling enough to overcome the perceived risk and negative inertia that stood as barriers to adoption. As a result, Object Databases occupy only a low-volume niche in the marketplace, although companies did benefit as various relational database vendors incorporated a subset of Object Database features into their products.
More recently, NoSql databases, one of the key technology enablers of Big Data, have risen into the spotlight. NoSql databases store data in a non-relational way that facilitates its distribution across many servers. While Object Databases began life as commercial products, NoSql databases were the brainchildren of Google, Yahoo, Amazon, and others – influential, technology-driven companies that shared the challenge of reliably storing and analyzing enormous volumes of distributed data. Their scalability requirements forced them to challenge conventional data storage wisdom in order to gain competitive advantage. Their willingness to incubate and open source their technology provided a means for others to better assess the risk of adoption. Their success in the market also provided a clearer barometer for assessing the value of rethinking one’s own technology strategies. Not only do the open-source NoSql databases that have grown from Google, Yahoo, and Amazon’s efforts provide new and unique benefits, but they have the momentum that the original Object Databases lacked – industry champions and substantial developer mindshare.
By raising expectations for what database software should provide with respect to performance, reliability, and cost of ownership, NoSql databases have made an important contribution to the Big Data ecosystem, enabling new possibilities for increasing revenue, lowering operational costs, and developing greater insight and agility. These new possibilities can be explored with questions such as:
- What if I could dramatically reduce my database software costs by paying only for production support (i.e. the software itself is free)?
- What if I could dramatically reduce my hardware costs by replacing a few heavy-weight, highly-specialized machines with commodity hardware?
- What if my data storage solution was immune to one or even two simultaneous hardware failures?
- What if I could take care of a hardware failure on Monday morning instead of Saturday night?
- What if my business decisions were based on more complete and more granular data? Could we gain competitive advantage?
In conclusion, technology can (and should!) drive business strategy by changing the game and forcing companies to adapt to remain relevant. At this point, it seems clear that Big Data, and NoSql database technology specifically, is not simply a “shiny thing” or a buzzword-heavy fad, so perhaps it is time to ask, “how could my company benefit from the new Big Data storage possibilities that now exist?”
-Andy Murray (firstname.lastname@example.org)