However, big data doesn't automatically mean better data and it's important to remember that there are reasons that we don't yet record everything about everything: because a lot of that data would be junk.
This is even more important when it comes to analytics in a company that hasn't really considered the applicability of its data before, or hasn't organized it in a manner that makes it all usable. Just because it has the hair color of every customer, doesn't mean it's going to be relevant to the business in the future. Figuring out what's important can be even harder if the data is largely unmanaged, with people either unable to give you access to all of it, or if they aren't able to help you understand it correctly.
There could even be petty office politics to circumnavigate in obtaining the data you need. Indeed some data may be siloed off from everything else, due to being part of an entirely different application, project, or team of individuals. Whether that data is worth retrieval is a major point to consider, as it's important not to open the flood gates and bundle all of it together just for the sake of it. Not all data is created equally and you could muddy what potential results you could find with more selective data usage.
So the focus should be about finding the right data, rather than all of the data, and several factors can come into play when doing that, much of which will be based on how well you know the business in question. As well as throwing out data that may be largely useless, other factors to consider include speed and cost, where ancient data centers may take a long time -- and potentially cost a lot of money depending on the impact of that transfer on the enterprise -- to access that information. If a quick turnaround is required, that data could be less useful than less applicable, but more readily available information.
Of course that won't always be the case, but it is worth bearing in mind when dealing with companies that are less familiar with cross-application data gathering.
Ultimately any company with its data hoarded away in different departments, with different access levels and antiquated hardware is going to conclude your interaction with a lot of recommendations for improvements, way beyond the results of the analytics you've offered them, but it's important to remember that whatever their drawbacks, be it too much, too little or just not the right type of data, the most important thing is to focus on the most actionable data.
The information that can really make a difference and most fits the customer's needs.