Ever since my early teens I’ve been a big basketball fan. I played football when I was young, but a playground accident (in which I broke both my arms at the same time) meant full-contact sports were off the cards for an extended period. During my recovery, I found basketball, and I never looked back. This was also right in the midst of the Michael Jordan-era – Charlotte Hornets jerseys were everywhere, Shaquille O’Neal was smashing backboards on TV. Basketball was blowing up in the early nineties, and like many passions of our formative years, it took hold and has stayed with me ever since.
One aspect that really captured my imagination was statistics. I collected NBA cards, poured over the numbers and info on each one. I developed an encyclopedic knowledge of useless facts about players and their outputs – you wanna’ know who had the best three-point field goal percentage in the 1992-93 season – I got you. Need to know the career averages of Bill Wennington? Right here. I wasn’t alone in this, there were a heap of people more informed and more detail-oriented than I, but what I didn’t know was that that very passion, that interest in obscure details and numbers, would one day change the very way the game was played.
Evolution Through Analysis
At the 2012 Sloan Sports Analytics Conference, cartographer Kirk Goldsberry gave a presentation on what he called CourtVision, an advanced basketball analytics system he’d put together in his spare time. Goldsberry had worked out how to extract data from ESPN’s shot charts – which showed where each player had made and missed shots from during each game – and he’d put all that data into a comprehensive set for each individual. He’d mapped every shot taken in every NBA game from 2006 to 2011, a huge data bank which, when filtered down to specific players, highlighted tendencies, weaknesses and strengths.
Basic field goal percentage data was nothing new – as noted earlier, any kid intoxicated by the smell of a freshly opened pack of basketball cards had some level of similar insight, but Goldsberry had taken it to the next level. He’d sought to show why this data was important, how it could be actioned. And as he presented, an audience full of NBA owners all sat forward in their seats.
Data analytics in sports has become the “in” thing in recent times. Growing from the success of Billy Beane’s “Moneyball”, analytics is now big business – virtually every major sports team now employs some level of data analysis in their preparation and evaluation process. And it makes sense – winning is everything in professional sports. More than pride or showmanship, it’s winning that makes money for pro athletes. Careers depend on it, clubs rely on the ability to perform. Winning teams get better attendances, more TV coverage, more success as a business overall. And it is just that – business. While it’s sports and it may not seem so different from your local leagues, where participation in itself is seen as a level of success, professional sport is a massive industry, and winning is a fundamental requirement. You’re either winning now or you have a plan to win in future. Or you’re done. With so much riding on the result, every little bit matters, every advantage you can get helps – if deflating the ball by one p.s.i can provide some tiny advantage, you best believe someone will try it.
With every detail under so much scrutiny, professional sports teams need to get things right. You could fly blind, stick with the way things have always been done – rely on your gut instinct, as many traditionalists still uphold. But the fact of the matter is data has become a critical part of modern pro sports. Numbers don’t lie, statistics are fact, and while it takes more than mere numbers to build any actionable insights from the info, used well, data can unlock the secrets that lead to that one goal – winning.
Data vs Instinct
Goldsberry’s formulas, or variations of them, have been adopted by players and coaches all across the NBA. The actual results of this are difficult to definitvely pin down, fuelling critics of the advanced statistics and data approach. Some, like TNT commentator and NBA Hall of Famer Charles Barkley, have come out strong with their opposition:
All these guys who run these organizations who talk about analytics, they have one thing in common — they’re a bunch of guys who have never played the game, and they never got the girls in high school, and they just want to get in the game.”
Barkley’s view is simple – all the numbers and all the data have not yet lead to a team winning a championship. And he’s right, but still, many clear winners have emerged.
Shane Battier defined his career by being a defensive specialist, someone who’s sole aim was taking on unglamorous task of shutting down opposition scoring threats. Battier was also an analytics advocate, someone who’d seen the power of numbers and had been using similar statistical correlations for some time. Battier became renowned for his success in stopping or slowing the game’s biggest stars, most notably Kobe Bryant. What Battier had determined with Bryant was that he was no where near as efficient when he shot from particular sections of the floor – so rather than work to stop Bryant, as such, Battier tried to keep Bryant out of his hot spots and shepherd him into taking bad shots. The tactic was a success, but one which isn’t necessarily quantified in the box score.
This sort of basic extrapolation of the data highlights the subtleties of utilising performance statistics as a predictor of successful behaviour. The data itself was never going to alter the nature of the game, but the accumulation of those subtle complexities, when used and applied in the right way, can sway the outcome and deliver results. The problem is that you a) need to know the right data to analyse and action, and b) need the right personnel to action it. Those two variables are what leads to data being seen as an inexact science – generally, it’s not a case of 1 + 1 = 2 – it’s more like 1 (in the right scenario with the right preparation) + 1 (with the correct understanding of the specifics of the moment) = 2. This is where there’s some truth to the old ‘go with your gut’ way of thinking – you need people who can ‘go with their gut’, but that gut needs to be informed and to understand the variables of overall success.
For instance, let’s say you have the ball and your team’s down by one with only seconds remaining and you’re rushing up court for the last play when you spot your teammate open for a shot on your left. An informed, analytical, mind will know how good that shot is, how good a shooter that player is at this stage of the game. Through understanding the shot charts, like Goldsberry’s CourtVision stats, the informed player can make a smarter decision and either execute or switch the play, and that quick thinking can win or lose the game. Such interpretation is both gut and analytics, and that’s more likely where you’ll see success in the world of data – human interpretation layered over informed insights. One without the other is an inferior approach.
New Ways of Working Require New Ways of Thinking
This is an important distinction in the intersection of big data and human analysis. Right now, the business world is trying to understand the implications of all this new data we’ve been given access to. The proliferation of social media has fed an explosion of online tracking and data systems and most business haven’t yet been able to get a grasp on what all this new information means, where it might lead. We know it’s important – if professional sports teams are effectively entrusting their success to the numbers, then it’s surely valuable – but because there are so many variables, because it isn’t so black and white, many are opting to stick with the ‘go with your gut’ approach, the ‘we’ve done it this way for years’ ethos.
So a heap of people on Facebook click ‘Like’ – so what?”
Established mindsets pose the biggest challenge to the possibilities of data, because it’s hard to see the logic when we’ve never been asked to look at things from a wider view. As with the quote above, a single person clicking ‘Like’ on your Facebook business page is virtually meaningless in the larger scheme. But we’re not talking about one thing. Often we go looking for simplicity because it’s what makes us comfortable, it’s logic we’re familiar with. But new ways of working require new ways of thinking, and we need to break out of what we know in order to break through.
Here’s an example in practise:
- Person A has 500,000 followers on Twitter. Person B has only 5,000.
- Person A has followed a heap of people and gained these followers over time by collecting as many people as possible, following whoever will follow back, actively seeking to up their follower count at every opportunity. Person B has never focussed on followers, but has instead focussed on community and having genuine interactions with the people to whom she’s connected.
- Person A has a Klout score of 55. Klout score, whether you agree with it or not, is an indicative measure of how many interactions a person has within their community, how many times they’re mentioned, the impact of their actual conversations. Person B has a Klout score of 75. This would suggest that despite Person B only having 1% of Person A’s following, Person B is actually more influential in their community and more likely to have her message reach a wider audience.
Knowing the above details, I’d be willing to be large sums of money that most people would still pick Person A and his 500,000 followers to be their brand ambassador over Person B. Because Person A has the biggest reach. The fact that they’re not listening to him is largely irrelevant – because we’re used to seeing things as we know them. What we know is that reaching more people is better – years of marketing and advertising theory has taught us this. We know that the chance of reaching 500,000 is better than reaching 5,000, because the audience is so much bigger. So what if not all of them are listening to Person A – even if you can reach 1% you’re still beating Person B, right? Even though, through the logic detailed above, we can see that partnering with Person B is probably more likely to generate better results, the majority of people will still go with what they know. The unknown is exactly that, and despite our data getting more informed, our approach isn’t quite there yet.
Data Analysis and the Evolution of Expectation
So going back to Goldsberry’s CourtVision stats – what if there was a way to correlate that same info, but for people who are buying or are interested in your products? What if, rather than shots made and attempted, you were looking at actions taken online – pages liked, interests listed, relationships. One of those things in isolation is nothing – someone who buys your stuff also happens to like Nirvana, so what? But what if, like Goldsberry, you could collect a wide set of data, a range of actions and preferences and map those on a chart which suggested that a person who undertakes certain, specific actions is highly likely to be interested in your stuff? You can do this. You can do this right now with Facebook data and Twitter info – you can correlate all the info from your pages and fans and you can build your own data sets that will map out the people most likely to be interested in buying from you. The trick is in finding the right data, the data you need.
For instance, correlating all the data from all the people who’ve liked your page might not be beneficial, because many people like pages for different reasons – they might be friends or family, they might have done so to enter a competition. Those people are going to skew your data, because they’re not the people who are most likely to buy. But you can narrow it down, specifically, to people who’ve made a purchase, to people who’ve interacted with your content. You can choose the specific info, most indicative of your typical customers, then build your datasets based on that. As noted recently, Facebook likes can very accurately indicate a person’s personality or leanings, when applied on a wide enough scale – those findings are the perfect business-case for conducting your own analysis and working out your own most relevant audience. Once you know this, you can target your marketing accordingly, you can focus your questions based on the queries amongst this sub-set, you can calibrate your focus around expanding your reach to people similar to this, people with the highest probability of being actual paying customers.
But that’s not broadcast reach, right? That’s not hitting the widest audience possible, which, as we know – as we’ve learned – is how to succeed and sell more stuff. And of course, that may well be the case – focus your dataset wrong or too narrow and you could miss out on an entire market of other buyer personas you’re not catering for by honing in on one group. Narrowing focus is a risk, and that risk is going to enflame oppositional forces, the old-school chiefs who know how things are done. This is the challenge of being an innovator, and has always been the challenge. You’re presenting a new way of thinking, and people aren’t necessarily going to like it. When you’ve achieved success by doing things a certain way, do you appreciate it when someone new comes in and suggests something different? No. Because you’ve done it, you’ve got the runs on the board, you have the experience, and experience is concrete. You know what works. Social media and big data are new, they’re different, and they’ve got a lot to prove – this means you, by extension, digital marketers have a lot to prove also.
But it can be done. The stats and figures can be located and correlated, you can work out the most minute and specific details about your target customers, and those details will inform the future of your audience approach. As communications become more individual, as more and more people grow-up online and develop their interactive and communicative skills via social media platforms, people are also growing to expect their voices will be heard. This is what social media is about, empowering people by giving everyone a voice – the brands respect and listen to those individual voices will advance and move ahead, in-line with customer expectation. Targeted advertising, for example, is becoming so specific that it’s scary – but to the next generation it won’t be scary, it’ll be how it’s always been. Brands responding in real-time will be standard, individual preferences will orchestrate the detail of each person’s media experience. What we know and have always known is evolving, whether we like it or not.
The possibilities of big data are amazing, the breadth of social media is hard to get your head around. But what we can say for sure is that people’s experiences and expectations are moving away from what we’ve always known. The businesses that can move with it, will.