The entire purpose of analytics is to provide us with information that allows us to make the best decision possible. Unfortunately, there are times when the data does not make the path forward a clear one. These difficult decisions can take time to sort out between stakeholders.
For example, let's say you are in charge of the marketing department of an organization, and you are testing out a specific new campaign. You select a subset of locations to present the advertisement, then track the results. At the end of the trial, you discover that for every dollar spent on the marketing campaign, $4.50 was generated in revenue for your company. Is the campaign worth scaling up?
These types of decisions arise often in the course of using analytics. Very rarely is there an obvious answer as to how to move forward. It's clear that the trial campaign worked, but did it work well enough to justify a more significant investment? On one hand, it made money! $4.50 per dollar is a solid return, and not likely to be viewed as a terrible decision come your performance review (assuming that when you scale it up it shows similar results).
On the other hand, there is not only the risk that it won't perform as well with a larger sample, you also need to consider how your marketing budget can be better spent. What if you have existing marketing strategies that generate $6 in revenue per dollar invested? Now the $4.50 return doesn't look so good.
Difficulty in making a decision isn't just emotionally taxing; it costs your company time and money. Additionally, there are political complications to going against the preferences of one decision maker to satisfy another. How we can avoid the lengthy process of making hard decisions?
Discretification! Don't bother googling it, for you'll come up empty. However don't let its absence from the dictionary prevent you from becoming more decisive. Before we touch on what "discretification" means, let's revisit two basic types of data.
Discrete vs Continuous Data
You have no doubt heard these terms before, however let's refresh your memory just in case.
Discrete data simply fit into distinct categories. A good example of this type of data is days of the week. Even though 11:59:59 pm Monday and 12:00:00 am Tuesday are one second away from each other, those seconds are in distinctly different days. Other examples include blood type, country, and genre.
Continuous data is a measurement that flows from one value to the next. Temperature is a nice example, for while some people may consider 90 degrees Fahrenheit to be "Hot" and others consider 20 degrees to be "Cold", there is no categorical difference between temperature readings.
An example that is forever present in business is of course . . . money! At the end of the day, no matter what decision we make comes down to the bottom line, which is inevitably a continuous measurement. In general, it is easier to make decisions from discrete data. Each category can lead to a specific course of action, requiring fewer discussions.
Unfortunately, because we are evaluating return on our investment most of the time, we are likely to be dealing with continuous data when making decisions. Changing the way we look at the data would certainly address that issue, which brings us to . . .
The Meaning of "Discretification"
It's time for the punchline.
"Discretification" is the mechanism of making continuous data discrete.
If you really grasp the concept, you may be thinking "Wait a minute, the type of data we are collecting is discrete in and of itself! Data can EITHER be discrete OR continuous, it can't be both!" You would be correct.
But what if we manually selected values along that continuous measurement, and declared them to be in a specific category? For instance, if we declare 72.0 degrees and greater to be "Hot", 35.0 - 71.9 degrees to be "Moderate", and anything lower than 35.0 degrees to be "Cold", we have "discretified" temperature! Our readings that were once continuous now fit into distinct categories.
So, where we do we draw the boundaries for these categories? What makes 35.0 degrees "Cold" and 35.1 degrees "Moderate"? At is at this juncture that the TRUE decision is being made. The beauty of approaching the challenge in this manner is that it is data-centric, not concept-centric. Let's walk through our marketing example first without using discretification, then with it.
You have completed your study of a new marketing campaign, and it returned $4.50 in revenue per dollar spent. The campaign was the brainchild of your boss, the marketing director, and they are proud of the campaign itself and the results.
A meeting is called between you, the marketing director, and the VP of Marketing to determine whether or not the campaign warrants an immediate investment. Your director urges the VP to give the campaign the green light; it's their project, their blood, sweat and tears showed it would help the organization. The VP is under pressure to deliver greater margins on their marketing expenses and given a $6 return in other avenues, they are leaning towards directing their funds elsewhere.
Ultimately, they decide to leave it up to you. If you go ahead with the project, you risk the VP having the impression you aren't making the decision for the good of the company, and that you are siding with your manager. Veto the idea, and you've damaged your relationship with the person who has the most say as to whether you get promoted (not to mention working with them every day).
Understandably, you want to be as careful as possible. You take a week or two to do research on other marketing ideas that could be pursued with the capital if it's not invested on scaling up the trial campaign. Other professionals you know weigh in as well in an effort to help you make the best possible choice.
You end up deciding that . . .does it really matter? You are damaging a relationship with a key stakeholder either way, and the process of coming to your decision took time, and time is money.
Let's now see how that story could have gone if you discretified your results.
BEFORE you move forward with the trial, you call a meeting between you, your manager, and the VP of marketing. At the forefront of the agenda is determining the minimum return that you all agree would need to be achieved to justify scaling up the campaign. The VP pushes for $6.50, to ensure it shows the department is improving profitability. Your manager calmly points to the location being used for the trial, and that is likely to perform worse than the rest of the population. Your manager is able to negotiate the VP all the way down to $5.75 as the cut-off point.
No doubt, you can see where this is going. The results come back in at $4.50 per dollar spent, well below the minimum. You simply send an email notifying the key stakeholders of the facts, and based on the PRIOR agreement cease pursuit of the new marketing tactic. No extra time needed to deliberate.
While your manager may be disappointed, this decision is less a reflection of them, their ideas, or their work. It's based solely on the data that was collected. This is what is meant by being data-centric versus concept-centric. This spares everyone involved from political complications. It's an objective decision that has the support of all key stakeholders.
There's an important variable in this process that is crucial to not only saving you and your organization time and resources, it also helps to IMPROVE the decision itself!
Timing is Everything
Perhaps you've heard of the famous experiment where they offered children one cookie now, or two cookies in fifteen minutes. The children that chose to fight their impulse to go for the immediate reward were shown to be more likely to succeed in their careers.
It's no secret delayed gratification works to our benefit. But what does it have to do with discretifying our data to assist our decision making?
The fundamental concept that ties this approach with delayed gratification is time. Several years ago, scientists created a function that captures how people value different outcomes based on how long it would take for the outcomes to occur:
The take-away here is that people tend to incorrectly discount the value of an outcome the further away it's expected to affect them. In the context of our example, when the results come in from the study and a decision must be made, the impact of that choice will be in the immediate future. According to the chart, if you are weighing whether to move forward with the new marketing tactic you are more likely to give it the green light. You are expecting to see rewards from it soon, meaning you are giving more weight to that outcome than taking time to find better marketing avenues worth pursuing, whose benefit won't be felt for some time.
By essentially making the decision by drawing the boundaries as to how you will discretify your data AHEAD OF TIME, you are less prone to this bias. Even if you end up green-lighting the campaign, you will have done so well before anticipating immediate impacts. Being less prone to this bias means more objectivity and thus, better decision making.
Let's finally look at a lighter example that exemplifies this concept. In the hit TV show "Deal or No Deal" contestants are given a briefcase with an undisclosed amount of money in it. They are constantly given choices as to whether they want to sell their briefcase (which could have anywhere from $.01 to $1,000,000), based on the values of the cases in front of them.
A simplified example: there are two briefcases left, one of which they own. One of the cases is worth $10, the other is worth $200,000, but the contestant does not know which is which. "The Banker" offers them $95,000 for their case. Do they take the deal?
On the show much of the entertainment comes from watching the contestants dramatically consider their options and take their friends' and family's advice. Because these contestants will receive the money immediately, the prospect of winning those prizes impacts their decision making. Perhaps they would be better off taking a significant deal, but they are thinking about how great it would be if their case had the unlikely million dollars? Maybe it's worth going for a larger payday, but they are concerned about making the wrong decision and leaving money on the table in front of millions of people?
These thoughts that distract us from objective decision making are likely to influence us when we are closer to feeling the impact of those decisions. If the contestants walked into the gameshow knowing the amount they would agree to and nothing less, it would protect them from losing objectivity.
Putting It All Together
We now know not only must we discretify our data to make our decision making easier, the time and manner in which it is done is essential. Let's recap how that process should work.
1. Identify the target
Every organization has specific, measurable goals (or it should). The data that is informing whatever decision you are making should put your team in position to hit those goals. Anything less isn't worth your precious resources.
2. Obtain buy-in from key stakeholders ahead of time
The most important decisions are going to be the most scrutinized. Having explicit communication with the executive team (or whoever necessary) as to how and why a decision is being made not only puts everyone on the same page, it protects you if things go south.
3. Gather the data
You know your target, now it's time to let the data do the talking.
4. Analyze and present the results
Fill in your team on what was found, reminding them on the targets you had set earlier.
5. Execute the decision
This is now the easiest part. Everyone agreed on the actions to be taken given the insights gained from your data gathering, all that's left is simply to follow through.
If prior to going on "Deal or No Deal", a contestant gave serious thought as to an amount of money that would significantly change their lives, they could discretify potential deals offered to them. They wouldn't be swayed by the spotlight or by the audience, friends and family urging them to go one way or another. They would calmly open cases, immediately turn down deals that were less than their target, and instantly take a deal when it was sufficient.
That fast, objective, unbiased, unemotional decision making process would make for a pretty boring episode. Maybe that's why there are no reality TV shows based on decisive data analysts.