I have been told data is the new currency by many people. I’m not sure how true that is, but like currency, data is useless without something to use it for.

As an example; 17,1.5,1,3,1.83. What is it?

It’s certainly data, but data about what?

Is it from a census? A line from a bank statement? A super weak password? Nuclear launch codes?!!

None of the above, not quite big data either but this small amount of data allowed me to make an informed decision recently and ultimately that’s what we want to be able to do from any kind of data.

“Should I remain in this line to pay a utility bill or should I move on with the rest of my busy schedule?” – That’s the question I asked myself.

Here is how I got an answer:

I observed the number of people in the line ahead of me was 17. Using my mobile phone, I recorded how long it took the next 3 persons in the line to complete their transactions.

I recorded 1.5 minutes, 1 minute and 3 minutes respectively for an average of 1.83 minutes per person.

Based on ~18% (3 as a percentage of 17) of the people ahead of me I now had an approximation of how long to expect each person’s transactions is going to take.

17 less 3 leaves 14 which when multiplied by 1.83 suggests I could be waiting for more than 25 minutes!

Clearly not acceptable so I left. I felt comfortable making that decision because of the data I was able to gather right there in the line with some added processing.

Now this process was in no way flawless. For example, how can I be certain the average time taken to complete a transaction would in fact remain around 1.83 minutes?

What if someone tries to pay in one cent coins or the card machine breaks down? Some people may also become frustrated and exit the line as I did changing the variables involved.

It would be reckless to claim that every transaction ahead of me will take 1.83 minutes but that’s the beauty of this use case, I did not need exact figures, I just needed to an estimate.

Being able to calculate one based on facts is more likely to be reliable than a random figure. Furthermore, had I decided to remain in the line I would have done so with my mind prepared to face at least 25 minutes of torture.

There may be an opportunity for some math wizardry in here (something about the rate water flows out a pipe?) but we will save that for another day.

The entire ordeal lasted roughly 5 minutes before I left the building.

Of course I could have just looked at the size of the line and left saving even more time however, 17 people really should not take 25 minutes or more to make a simple payment.

The length of time taken of course depends depends heavily on the customers and the cashier, which brings up another point:

While I consider this small data, in the sense that I only looked at 17 people; aggregated, this information may be useful to the institution.

How long are transactions taking on average each day? What type of transactions take the longest and why? Which cashiers seem to be operating the slowest?

All real problems this method of sampling could help identify if applied at the organisation level. Organizations able to collect this data can use it to optimize their processes and provide a much better customer experience.

Remember, happy customers bring brand loyalty, in the case of utility services, less slander of your brand on social media and more bills paid on time.

So what do you think? Is small data worth… the time?

Leave a comment and share your thoughts!