What does a caveman hunting wooly mammoths and an entrepreneur launching a startup have in common? They both depend on data to survive.
While we often treat data as this hot new thing, it’s actually as old and primal as human beings. Animal tracks, weather patterns, geographical features, memory of migration patterns, smells and stories were all rich sources of ‘data’ for the hunter gatherers.
They survived if they asked their data the right questions. Businesses that wish to thrive must do the same.
The difference today is the ‘bigness’ of data. Hunter gatherers had a limited amount of data to pull from at any moment. They distinguished the useful information from useless, or they made potentially fatal decisions.
>See also: Beneath big data: building the unbreakable
People today create trillions of gigabytes of data, which makes it that much harder to sift the gems from the junk. But the right questions can lead businesses to examine the right data.
The digital universe is expected to reach 35 trillion gigabytes by 2020. If only 20% of the data accounts for 80% of its value, that 20% is still an overwhelming amount. Companies that thrive will find that 20% by training their questions on three main areas: revenue, cost and customer identity.
The data trifecta
Questions vary from business to business and objective to objective but revenue, cost and customer identity are going to cut across all companies.
The idea here is not to focus on cookie cutter metrics that accounting can answer, but rather to ask a series of questions that produce useful knowledge.
It’s common to find lots of 80/20 scenarios when analysing data this way. For example, 20% of customers probably account for 80% of revenue, and 80% of costs (returns, customer service, implementation issues, etc.) are probably linked to 20% of customers.
So who’s who? What do the high revenue clients have in common (demographics, affinities, verticals for B2Bs)? What do the pain-in-the-butt, expensive clients have in common?
The intention here is to figure out how businesses can replicate the good and eliminate the bad. Companies tend to think of metrics in black-and-white, more-is-better terms. Customer growth is good; attrition is bad.
But what if growth is only coming from the worst kinds of customers that make life hell for employees? What if social media growth is really driven by click farmers that make it more expensive to reach people who care about the brand?
Questions about revenue, cost and identity force companies to confront questions about quality. Data is not created equal, and nor are revenue streams, costs and customers.
The right questions
A goal-setting framework called Objectives and Key Results (OKRs) leads to questions about revenue, cost and customer identity. It keeps companies honest about data use, and helps businesses trying to sift the valuable 20% of data from the mediocre whole.
If a business’s objective is to launch, market and sell product X, it comes up with three key Results that will measure how well it achieves it: 1,000 customers contacted, $1 million in revenue booked and average customer satisfaction rating of 4.5/5.0.
If the company hasn’t launched any products before, it doesn’t have much data to interrogate. It will be trial and error the first time around. However, if it does have data from past product launches, there’s a lot it can ask.
Last time if contacted customers, how many returned its messages? Did calls, emails, LinkedIn messages, etc. get more or less responses? How many converted, and what made members of that cohort different from those who didn’t convert or respond? What do the converts like, what’s their income level, education, credit score, etc.? Data can help identify 1,000 people with the highest likelihood of buying.
For $1 million booked, company should look at what each customer historically spends, especially if price is going to vary across different user bases, product packages and other variables. The business wants to know what customers consider a fair price so it can either a) set a flat price likely to generate $1 million, or b) go into negotiations knowing what outcome is optimal.
With customer satisfaction, ask questions about the historical patterns. What features or outcomes were the source of five star reviews? Which generated sub-4.5 reviews? What extra costs would incur if you built out and administered a knowledge base to stem negative reviews from customers who had problems getting their product setup?
>See also: Is big data dead? The rise of smart data
OKRs will discipline a company’s relationship with data. Hypothetical scenarios like this also show what data it should be collecting if it isn’t already.
If you run a business, or hunt woolly mammoths, some data is more important to you than other data. The importance of some data changes over time. Purchases histories stay valuable for years and maybe even gain value as the patterns crystallise.
On the other hand, data on how people use a website is going to become useless as soon as the site updates, so the data should be discarded and not used again.
It feels exciting to be on the front edge of the data ‘explosion’, ‘revolution’ or whatever you’d like to call it. Companies will make soberer decisions and ask better questions if they remember that they’re using the same mechanisms that turned woolly mammoths into dinner. Data isn’t sacred. Great questions might be.
Sourced from Korey Lee, CIO at SumAll