The Problem With Outsourcing Data Quality

The Problem With Outsourcing Data Quality

Aug 17, 2012 by in Data Quality

There are many companies who now offer the ability to outsource your data quality woes away. The idea being that perhaps complex data quality processing or cleansing can be facilitated by a more specialist or cost-effective option elsewhere, often overseas.

On paper this may look like sound business sense. Most organisations employ some kind of outsourcing arrangement these days, especially for IT services, so why not bundle data quality into the deal?

Now, before I’m flamed, let me say for the record I’m a massive fan of outsourcing, it has certainly helped me find specialist staff across the globe who have helped me solve a range of resourcing challenges for my own micro-business but I have one primary concern when it comes to data quality outsourcing and this is the issue of status quo preservation.

Outsourcing data quality is normally associated with the mechanical processing and transformation of data. Perhaps you’re deduplicating customer records or validating batches of equipment codes. Whatever the process, by moving this activity outside of your primary business operations you’re effectively handing out over control to a 3rd party.

In my experience, when this happens the process invariably gets forgotten and data quality is one activity that should never be forgotten. Data quality is a continuous improvement activity, with a focus on improving processes, not just data. By outsourcing data quality you simply persist the status quo and that doesn’t sit right with me.

Yes, it can make sense to outsource data warehousing administration, ETL management and even software development but something that involves quality should surely be firmly in-house, tightly managed and constantly reviewed.

Some data quality activities of course can be outsourced, such as monitoring, measuring, alerts, rule building and anything else that is peripheral to the core data quality function but in my opinion you should never outsource an entire quality process, data or otherwise.

But what do you think? Welcome your views.

Read more posts by Dylan Jones on
the Data Roundtable.

Tags:

One Response to “The Problem With Outsourcing Data Quality”

  1. Olivier Kenji Mathurin

    Aug 19, 2012

    Hi Dylan,
    Great post on a great topic! As Data Quality related processes are in most case not a core business process, one can be tempted to outsource it, as you highlighted. And I totally share your view that keeping core Data Quality processes in house is vital to a firm willing to eliminate the business issues related to the quality of the data.
    I would however balance your position related to outsourcing data quality related activities. In capital markets and especially on the buy-side (banks), outsourcing of data activities for investment data (securities, prices, corporate actions) is pretty common and handled by Business Services Providers (BSP), often subsidiaries of financial firms who wants to leverage an expertise they have built, and that other banks see as non core processes: capture, matching, cleansing, enrichment, validation. These activities are however seen as critical for the business to run, but the expertise required – due to increasing complexity, volumes, regulations, multiple countries – is very expensive to maintain in-house.
    Outsourcing can be a solution when 1) the processes are standardized 2) strict SLA are put in place 3) outsourced activities can be audited (which data was matched using which set of rules, etc.) and … 4) somebody in the firms is ensuring SLA is respected and is able to trace back the origin of the faulty data when a business process fails for claim purpose.
    In brief to make outsourcing successful, you already need some level of maturity in the firm. And it is business related, not IT related. Outsourcing wont be a silver bullet to solve your lack of data quality focus…

    Reply to this comment

Leave a Reply