“Learn to code.” Around the world, it has been a familiar refrain. And the world has duly learned.

Time was that the standard in-school computer education was little more than word processing, spreadsheets and some basic programming. When I taught at Cal Poly in the ’80s, undergrads all had to take my class in basic computing. But today’s students, like my son who just graduated from UCLA, all take mandatory programmatic-centric classes. Economically motivated to churn out adults with the tools to write a future as a tech entrepreneur, education systems are transforming, from ones that major in “Office Products” and a smattering of Visual Basic, to ones that are more about Python or bust.

The federal government is pushing coding proficiency as an economic booster as well. In the past two years, for example, President Obama has lanuched a TechHire Initiative and a Computer Science for All program, which together saw billions of dollars in federal grants handed out to increase the hard coding skills of the workforce.

While these initiatives may indeed create a capable army of future Zuckerbergs, what it doesn’t do at all is serve the majority of businesses in America.

The lurch from employer-centric computing education to one driven by programming means the middle ground of tech skills – between basic computer proficiency and advanced code creation – has thinned out.

Schools just aren’t turning out people with the right skills.

I will wager that 99 percent of businesses in the U.S. don’t need anyone proficient in C++ or Java. The tech skills required by most employers are substantial but quite different:

1. The basics of a scripting language. Bash for Unix/Linux, JavaScript for web browsers, or Visual Basic for Microsoft Applications are simple coding skills that are easy to learn and valuable for workers across disciplines and levels. These skills allow you to automate tasks, promoting efficiency in manipulation and analysis.

For example, if you run a contest, you could write a simple script to determine if people who’ve entered the contest submitted their content to your site by the specified date. Looking up hundreds of users manually would be very tedious, but this scripting language know-how would make the process efficient.

2. Simple SQL commands. These commands are necessary to process raw data and turn it into information that you can analyze and apply.

Sure, the right people on your team should know how to code – but most of them should be writing spreadsheet macros and pivot tables to support your internal business processes, not agile algorithms for entrepreneurial endeavours. They should know the basics of HTML editing and how to set up folders and accounts with the correct security rights for your team. That’s what the bulk of businesses need from technology education.

3. Deductive reasoning skills. Being able to look at various pieces of data and draw a conclusion is probably the most valuable skill for any employee to have, and surprisingly it’s something that’s too often missing from otherwise technically advanced employees.

I had a statistics expert come in and present to me on our site’s traffic patterns. Unfortunately, he presented unhelpful raw data, with no deeper layers of context or meaning. When I asked him to break out which percentage of our audience hails from certain countries, how long they stay logged in, and what their top three on-site tasks were, he didn’t know where to start. While this expert had the basic computational skills, he lacked the analytics skills to understand the connections his data could prove.

Nowadays, the most in-demand coding skill in the world is not the one that empowers you to build a startup that boosts national GDP, it’s the ability to perform advanced analysis on data that creates rich meaning from raw numbers.

If education has overshot the mark when it comes to the real-world skills tech workers need, it falls to businesses themselves to pick up the slack. The next generation of employees can arm itself through independently-taken online courses, as long as employers make clear the kind of basic data analytics skills they expect their employees to have.

Gene Richardson is COO of Experts Exchange.