In the world of data scientists you’ll often hear the phrase GI-GO (like the soldier): garbage in, garbage out. That’s a little too politically correct, especially if you were on the receiving end of some of the poor data. We prefer crap in, crap out to be a more accurate depiction of the trench warfare that is data munging.
Crap in, crap out generally manages expectations for the non-data inclined. Too frequently I’ve seen mangey data sets handed to engineers under the false hope that the end result will yield whatever nugget of gold someone preconceived.
That’s just not true.
Data must first exist.
If a POS does not even capture transaction data (many legacy POS systems don’t for longer than a few weeks), then it’s not possible to run any sort of regression analysis on popular items. Asking someone to tell you what was popular last month if you’re using a Micros 3700 is impossible.
Data must have some sort of structure.
I need to know what I should be looking for and where it resides. If I’m looking for the name of products (let’s say chicken sandwich), is this field going to be in a column? Between some other demarcation? Even if that’s clear, it does me no good if the item’s become so abbreviated or so misspelled that not even your employees recognize it.
Yet an astounding number of brick and mortar merchants purchase point of sale systems and either a) are ignorant to the negligent data practices employed by the POS software, or b) never even bother to use the product as intended.
Let me put it to you like this:
Why would anyone spend thousands of dollars of their annual income on a car if they were only going to pop it into neutral and push it down the street?
But for some reason I’ll surely never understand, cost-conscious merchants easily fork over this much in annual POS expenditures without using the product fully.
Why should merchants care? In the event they’ve been living in a cave, the market is becoming more competitive. Consumers can now order similar – if not the exact same – goods from a larger number of online providers at cheaper prices.
And do you know why the prices at Amazon or a meal kit startup are lower?
Because they use data to optimize everything.
Merchants who are not optimizing their businesses are telling the customer, “Please pay for my inefficiencies.”
If brick and mortar merchants want to stay solvent they are going to need to start using data. Their competitors are using data whether merchants want to admit it, and there’s a reason why.
Before merchants even start thinking about using data they must recall the data scientist’s mantra: crap in, crap out. If the data is useless in its basic composition no amount of magic will remake it into something useful.
And this is where the POS dealer comes in.
Dealers need to teach merchants the fundamentals of setting up and using their POS systems to ensure that whatever data they do generate is actually useful. Sadly, in the cases of many legacy POS systems, this means dumping the POS altogether and starting anew.
As a dealer, if your POS provider is not investing in data replication (a term for moving data up into the cloud by default) or a cloud product line, you really need to be looking to carry another product ASAP.
Assuming the POS is cloud/has data replication and doesn’t simply flush the data every few weeks (hi Micros!), now you need to convey the value of using the POS as it’s intended.
Are employees sharing logins? Then how the hell are you going to objectively determine employee performance?
Are items abbreviated? I sure hope everyone on your staff remembers what cfopwqqni stood for when they need to figure out how to replicate its success.
The dealer is the only party that will be able to convey and cultivate this importance to a merchant.
Payments professionals are not getting into this level of granularity: they want the processing contract, will feed you a POS product if needed, but that’s the extent of their concerns.
Most venture-funded cloud ISVs are likewise solely focused on sales quotas and many have noticed they’ll do whatever it takes to make sales.
Using data is going to become crucial for survival. Sure, many merchants will rebut with, “Oh well people can just share logins because it takes too much time for them to login.” Really? So a 5% sales lift is not worth the extra second it takes an employee to log in? You’re not even that busy every hour of every day to justify shared logins anyhow.
Dealers can start coaching merchants up on these changes, and it will dovetail nicely with the direction the VAR market is heading anyhow: to be a VAR a dealer must sell knowledge. As cloud erodes installation, support and other revenue streams, good dealers are left with their ability to provide consultative value.
It’s something we intend to cover in more detail at a later date.