Data Analytics: To Outsource or Not to Outsource

Everyone knows that Big Data is the future and no one wants to be left behind. Positioning yourself on the forefront of the data revolution will be crucial to harnessing your data to make better business decisions.

But where is your data? How do you interpret it? What are the possibilities? How can you make such analyses affordable? Rather than hiring an expensive (and elusive) data scientist, consider outsourcing the job. There are many advantages of doing so.


1. Rapidly evolving analytics tools

On the analytics landscape, each passing day brings with it a novel tool to solve the analytics problem that you didn’t know that you had yesterday. Often, by the time you build a customized solution to a problem, someone has built a tool to solve it. Frequently with a more easily accessible interface than your solution has.

KissMETRICS, for example, helps e-commerce sites to collect and organize user events linked to email addresses to allow you to better know your customers, and to enable personalized re-targeting. Tableau helps you to see your data and build custom dashboards for consumption by different departments. Mention lets you know who is talking about you.

Will your in-house analytics team be on top of the next great tool?  Will they have the time to learn how to use it?


2. The data scientist and the corporate pyramid

It’s a new conundrum: determining the point of entry for the data scientist in the traditional corporate pyramid.

When your analytics are in-house, are they just analysts writing reports that no one ever reads and acts upon? Do they have the ear of the CEO? They are sitting on crucial business data—will they have the clout to make it heard at the top of the chain so that future business strategy can be based on cold hard facts?


3. The tool doesn’t a master make

Many technology companies are building amazing tools to bring analytics to the masses. Sounds like a great idea in theory. But does it pan out in the real world? Can you make a non-technical person a data analyst with a tool? Does accounting software eliminate the need for accountants? It doesn’t. As it is with analytics tools.

Data scienctists possess a varied skill set—and uncovering true actionable, value-producing insights requires analysts with a certain knack for diving into data without getting lost. In addition to having the talent to convey that information appropriately to the right people. There is often a need for a deeper drill-down into the specifics of the observed insight.

A colleague of mine often uses the term “Data Wrangler” for the work that he does.  When you want to narrow down on specifics, an untrained tool-user won’t be able to get to the bottom of the insight. A data scientist—a true data wrangler with low-level programming, parsing, AND mining skills—will.


4. Outsource to pare down the data

The real problem of big data is that it’s big. There is too much of it. A data mining session might produce 100 interesting trends and insights. Then what? 100 insights aren’t digestible.

The untrained analyst may get buried in the what-could-bes, the this-could-be-importants and all of the possible slices of the dataset.  A trained data scientist is experienced in separating the signal from the noise.  They will weave together the story of the data-driven, business decisions to make and will hold your hand through the process.


5. Do you even know what you have?

Many small- to medium-sized companies may not even be aware of the goldmine of consumer data that they are sitting on. Do you have customers that are navigating your website?  Then you have important data that will tell you where your business should evolve in the future.

It’s the conundrum of not having a physical store. When you can’t see the customers, it’s more involved to extract their shopping and browsing habits than it is when you sit at the counter and watch them saunter around the store.

Even with tools like Google Analytics, a trained data analyst is an asset to have access to.  The same arguments also apply to the startup space. You’re acquiring new users. Who are they? Where are they coming from and what are they doing? To grow with the demand and evolve into what users need, you need to be crunching the data. Hiring a permanent employee when you’re not sure what you need them to do can be risky.


6. Can you afford it?

A hired data scientist can run you up to $100,000 in salary alone. Does it have to be that way?  Many small- to medium-sized companies can get away with paying a modest set-up fee to begin their relationship with an outsourced data scientist, beginning with accessing and organizing their data. Then transitioning to an ongoing relationship with an outsourced data scientist that would cost them much less than an in-house employee.