Good stuff Prash. May I bring my favorite subject, being external reference data, into play? I have had good experience with using property data for this purpose. You are more certain about that it is a household if you know the address as a single-family house or you have an existing unit in a multi-occupancy building. Further, if the address is a nursing home or a university campus you should not try to group occupants into a household.

Great points Prash. I’d simply add that from a public sector perspective that household view is often critical. Whether it’s income generation ensuring the right level of taxes are paid, reducing fraud/error, local democracy ensuring everyone has the vote or the protection of the vulnerable that household composition is vital. Always a good moment when customers realise that the proverbial single view is more than just the citizen.

Hi Prash

Another pet topic of mine.

In South Africa individuals have a unique Identity number – many MDM projects rely on this number ofr matching (similar to a US Social Security number it should be unique to an individual). Not withstanding that poor data quality means that ID number is frequently not unique (or accurate) this cannot be used for householding.

So other shared attributes must be used for householding – address, etc.

from a census perspective the Statistics South Africa describe a household as a group of individuals who spend at least 4 noghts a week under the same roof – quite a challenge to model that 🙂

have a good weekend
Gary

Nice article Prashant. One telecomm client had an interesting case – households could be different – e.g., Service department is interested in Service Location; Billing department is interested in Billing Location – they called it economic household – e.g., you may be paying for your parents internet,tv and phone and you and your parents would be in the same economic household.

Really interesting stuff and if you add geocoding to this the use cases are endless.

Raja

Prash –

Great stuff. BTW, there are multiple forms of households, and an individual can be part of multiple households. For example, sometimes you are interested in people sharing an address, but sometimes you are interested in people who have financially dependent relationships or familial relationships. You may also find people who have multiple addresses, including children shared between divorced parents or folks who move between two addresses (e.g., college students, “snowbirds”). In a sense, household relationships are just like any other kind of relationship. Thanks for another great blog.

– Dennis

Household control of customers is a cool data setup for data-driven CRM in dialogue marketing. Purchasing patterns tell more about the household when adding up the individuals. Individuals are kept on own records and only intelligently viewed and utilized as statistical data about households. When individuals separate, decease, etc. the households are automatically updated. Here in Denmark, I mainly use the KVHX address key which is strong data match along with cleansing validation of the individuals. Works great. I understand other countries have more challenges. Great to learn how you match in other countries.

Hi Prashant,

Householding is an interesting topic indeed. Many of the customers still struggle to use to this useful concept in their MDM implementation lacking data quality. Many of the folks have presented some really interesting use cases indeed. I think integrating with external references such as census data or data from other public departments (viz.. IRS) etc periodically is a sure shot way to re-enforcing integrity of your house holding data periodically.

One another piece that I find fascinating is how with the explosion of social media, the concept of house holding has taken a “glamorous” turn. By analyzing the social media interactions within the house hold of family, friends and relatives one can try to “personalize” the results house hold experience. The integration of these two, which i believe is definitely happening somewhere in the retail industry represents a really amazing integration of concepts and technology.

-Vishu

Thx