The next time you tuck into a restaurant pizza, your enjoyment may owe as much to data crunching as to cooking.

The ingredients, the ambience, the pricing, the location - all these elements will have been finessed thanks to clever analysis of mounds of data.

This is because computing power and the range of available data sets have increased so much we can now make business decisions based on hard evidence, not gut instinct or guesswork.

Cocktail bars

"Casual dining" chain Pizza Hut Restaurants UK has been applying data analytics to its £60m ($94m) refurbishment programme, begun after private equity firm Rutland Partners bought the brand from US owner Yum! Brands in 2012.

Pizza Hut is just one of many firms embracing data analytics, which was pioneered in the UK by high street supermarket Tesco and its Clubcard loyalty scheme.

"We'd seen sales fall off over a number of years as more entrants came in to the marketplace between 2000 and 2010," says strategy director Andy Platt. "Operators like Nando's had done very well.

"Our range had become limited; we weren't as colourful or as exciting as our competitors; and the look and feel of our restaurants had fallen a little behind."

So the company added new pizza bases and toppings, jazzed up some of the restaurants with cocktail bars and warmer lighting schemes, and generally "funked" up its image to appeal to a younger clientele.

But how did they know any of this was working or whether the money was being well spent?

Signal from the noise

Increased sales would seem to be an obvious answer. But in the world of data analytics, the obvious answer isn't always the right one.

"It's easy to see results and take the wrong read of them," says Mr Platt. "An increase in sales could be to do with something else - seasonal summer holiday trade, for example - and not the change you made to the menu or the decor."

The answer to this conundrum, says Rupert Naylor, European boss of Applied Predictive Technologies (APT), is to run control experiments, much in the same way that they test for the efficacy of new drugs.

APT, which includes KFC, Beefeater and Brewer's Chain among its clients, did the data crunching for Pizza Hut Restaurants UK.

"We help people run tests on how to innovate and whether that innovation is working," says Mr Naylor. "We do a control test on restaurants that show similar behaviour as a baseline, then we strip out all the noise - the data that may have affected sales anyway - to get to the truth.

"It's a scientific, evidence-based approach."

An example of "noise" might be the sales growth that was occurring naturally in a restaurant because it was located in an area of increasing population, say. Your super new pizza topping might not have had anything to do with it.

But if you made the wrong assumption and rolled out the new menu across all your restaurants, regardless of their location or customer demographics, you could end up wasting a lot of money.

Waffle

So what did Pizza Hut UK learn from all this testing?

"We saw 40% growth in some of our really big investments, which was truly transformational," says Mr Platt. "Before we started, average customer spend was about £9, now it's around £11."

Across a chain of 250 restaurants each entertaining 1,000-2,000 customers a week, that all adds up.

"For Pizza Hut UK, the uplift in consumer spend from £9 to £11 through the use of data analytics is marginal, yet extremely important," says Greg Bromley, retail analyst at Conlumino.

"While the investment needed to leverage data analytics successfully is typically high, the long term benefits often have the potential to significantly outweigh this."

There were some quirky findings, too.

Waffles, it turned out, gave the restaurants a delightful aroma but cost so much to produce that they were never going to make any money.

"You have to keep a waffle iron hot all the time which takes a lot of energy," explains Mr Platt.

Smart meters were able to isolate individual appliances and work how much energy they were using, which helped reduce overheads further.

"We never stop trying to improve," he says.

Cognitive bias

Data analytics and continuous testing help us to draw the right conclusions from evidence - something that humans are not naturally good at.

We are prone to "cognitive bias", or wonky thinking, because we don't weigh up the evidence correctly or ignore it altogether, swayed by our prejudices, emotions and lack of logic.

"We often see patterns that aren't there and assume one thing caused another when it didn't," says Christian Chabot, chief executive of Tableau, a data analytics software company.

For example, one of its clients, Metro, a Singaporean retailer, used to offer its shoes in half sizes as well as full sizes because that was the done thing in the sector - it was assumed that's what customers wanted.

But a proper analysis and visualisation of the data revealed that customers rarely bought half sizes. Metro adjusted its stock accordingly, which led to reduced storage costs and less surplus stock.

"There's a huge thirst among organisations to run their operations better by consulting the facts properly," says Mr Chabot. "Before, we never had the right tools to do that - now we do."

And we have much more data to play with, too, including weather, demographics, traffic, purchasing patterns. mobile phone movements, social media profiles - literally hundreds of datasets that can be combined to reveal very detailed pictures of who we are, what we like and how much we're likely to spend.

Sandwich economics

In another example, APT client Subway, the sandwich chain, famously baulked at one franchisee's idea to sell a foot-long sandwich for $5.

It didn't seem to make commercial sense because the cost of the ingredients would wipe out most of the profit margin.

"We were able to show that it worked because more people came in and increased sales of other goods, like crisps, drinks and so on," says Mr Naylor.

"But our tests showed that it didn't work for every type of sandwich. The answers are always nuanced - never black and white."

In the era of data analytics, the retail mantra "know your customer" has become a lot easier to effect.

And if that makes for a tastier pizza, more power to their elbow.

Follow Matthew on Twitter: @matthew_wall