The 3P’s of data management

Referrals to “the 3 P’s” are common in management literature. A quick search on the internet gives quite a number of explanations for the abbreviation “3P”:

• Product, Price, Place
• Third Party (procurement, architecture, integration)
• Triple Play (television, telephone, internet access)
• Third Player
• Three Pointer (basketball)
• Proven, Probable, Possible (oil industry classification)
• People, Process, Product

Of these, “product, price, place” is the most commonly taught one in business school. Today, we introduce a slight twist on that one with regards to effective data management: policy, process, people. Let’s take a look at these 3P’s in more detail.

Data is everywhere; people get lost

Policy

Your data management policy really refers to the fact that, as a business, you have a plan about data. What this means is that you’ve decided on a — hopefully company-wide — approach to dealing with data. This starts with recognising that there is data and this data, more often than not is siloed and dispersed across many systems. The plan either involves tossing out all those systems and replacing them by one single entity. We’ve seen many attempts at doing so. It typically requires many years of implementation, process changes and, more often than not, frustration amongst users. A more modern approach would let systems in place in recognition than one-fits-all doesn’t really work. Heterogeneous IT landscapes consisting of ERP, CRM, SCM,… and other best-of-breed systems are becoming common because reporting tools allow users to gather data from multiple sources and combine them into actionable output. The new policy then requires a common understanding of what data you have, where it can be found and what it means.

Process

In order to then effectively leverage dispersed data, a process needs to be put in place to ensure the above mentioned mapping of the data. Quite like a GPS helps us find our way, the process should help us find data, find the right data, and make sure we understand what this data represents. Indeed, if we allow data to be managed by the direct users — who should be best placed to use it correctly — we must ensure that other people know exactly what that data represents. Take, for example a multinational with a European HQ and US entity. The US users may use as “purchase price” the price at which their entity buys goods, but if those goods were acquired through the HQ, that price is more likely the HQ purchase price plus inter-company pricing. For the US entity their definition of purchase price makes perfect sense, but for corporate reporting, a clear difference must be made between both definitions.

Managing the “data GPS” is complex and involves many people. To implement the process one can’t rely on paperwork alone. Just like GPS replaced maps and atlases, the process must be materialised through implementation of a tool. That tool would be a data catalog; a digitised process allowing stewards to manage and trace data and offering end users an easy and straightforward access to data stored across multiple systems. The data catalog, by its explanatory nature helps turn data into information.

People

A good friend of mine likes to say “a fool with a tool is still a fool”. He does this to emphasise that, however good a tool may be, the person handling the tool determines the result. Therefore, tools should comply with a number of requirements:

• the tool must fit the purpose: depending on the planned use, certain features will be required, others “nice-to-have” and even others will be unneeded (and therefore may render implementation and adoption only more complex). When selecting a tool, seeking advice from specialised consultants is usually a good idea.
• steep learning curves can kill a project before it leaves the starting blocks. Data stewards and end users will use the tool in very different ways; make sure that the training is adapted towards their needs.
• the tool should be efficient and effective. It should be easy to use, accessible, able to search swiftly to all and hardly require any explanation for data consumers (efficient). Your web search engine doesn’t come with a manual either! It should also be able to find the correct data (effective). Some searches may yield a multitude of possible answers. Similar to a search engine, answers should be classified according to likely importance for the person searching. This can be through ratings (by other users), division of the person searching, etc.

Adoption of the tool and continued usage are key for the success of your data management project. People are emotional beings and we must overcome the initial resistance to change when introducing a data catalog. This is achieved by the efficiency and effectiveness mentioned before. But research has shown that participative systems in which users remain engaged and are encouraged to provide feedback or participate in building out the content not only achieve faster adoption but maintain higher usage rates. User engagement is different for all. While many users are quite happy to just use the GPS, others are enticed to rate and leave comments (as per Google reviews or YELP for restaurants). Being able to contribute increases people’s sense of purpose and before you know it, you have a real data community.

As you may have guessed by now, while we build the tool, effectively helping to materialise the process, our true focus is on the people. People are the critical success factor of your data management project. Policy and process can be bought (consulting, tools) but the support of your users must be won. Therefore, whatever you do, make sure that your people will want to use the tool. While this can be as simple as “like that tool better”, emotional hurdles are harder to overcome than anything else. And it doesn’t depend on features alone, those in charge of the project will have the important task of picking the tool that will eventually incite the highest usage. As history has shown, you’d rather have picked the iPod than the Zune…

Therefore, in a spin on the classical “3P” expression, the 3P’s of data management really are People, People and People.