Why big data and analytics will not make you the next Apple
Are ‘big data’ and ‘analytics’ the latest fads?
In my view there is a little too much hype around ‘big data’ and the power of analytics to drive business growth and profitability. Can ‘big data’ and ‘analytics’ help you improve operations? Yes – they can help you better staff and manage your call centre or improve your supply chain or target your marketing better. Will ‘big data’ and analytics make you the next Apple? Highly unlikely. Why does that matter? Because whilst you are busy optimising operations someone else is inventing a future in which your optimised operations become redundant: think music, think publishing, think mobile phones. Let’s explore this further.
What are the root causes?
Years ago when I worked in the business planning & analysis team of a brand name drinks company I noticed something interesting. Managers could drill down and find out which particular markets had failed to make their numbers. Great – we know which business unit is underperforming. But why is this unit doing better or worse than expected? This is where the fun started. First, there were almost as many opinions as the people we asked. Second, there was no easy to work out whether the answers given by local management held any resemblance to the ‘truth’. When I dug further I found out that local management did not know why they had failed to make their numbers (revenues). When they were asked, the local managers simply came up with the most plausible story. So month end because a ritual where HQ asked questions and the local managers invented plausible stories.
Lesson 1: For any complex event there are a multitude of plausible answers and working out the ‘real reason’ is notoriously difficult.
Lesson 2: Without a sound grasp of the root causes it is difficult to formulate a sensible course of action to address the situation.
What do we do about it?
Lets assume that you have all the data and analysis has been done. You know there is a problem. What do you do about it? Is it easy to get all the actors – who have to play ball – to agree on the course of action to take? In my experience the answer is no: the more ‘strategic’ the issue at hand the more difficult it is to come to an agreement on a sensible course of action. Let’s take a look at the recent banking crisis. Did all the main actors (the western economies) come to a consensus on what course of action to take? No. Just take a look at the Euro crisis: why is it that the leaders of the EU cannot agree on the right course of action? First, because different people have different ideas about what constitutes the right course of action. Second, the right course of action from an objective perspective may simply not be viable from a political perspective.
Lesson 3: The more that is at stake the harder it is to get all the actors to agree to a single course of action and then act to play their part and execute that course of action
Lesson 4: If you are unable to act decisively and as a single unit then all the data, analysis and insight is worthless
Does it really tell you what you need to know to thrive in the future?
Larry Freed has written an article that resonates with me. He points out that you need to be clear on what you know and what you do not know. I’d say that you need to be clear about what data and analytics is telling you and what it is not telling you. Larry, talking about a website, asks do you know:
- “Why visitors come to your site (to research, to buy, to complete a transaction, to get product support, to learn more about your company before interacting with you through another touch point, etc.)?
- What influences visits to your site (a referral, a social media interaction, a failure to resolve an issue with a call centre, an advertisement, a news story, a previous affinity with your brand, etc.) and which customer acquisition sources result in traffic that is the mostly likely to convert?
- What visitors need from your website? How needs differ by population segment or other segmentation that is useful to your business—perhaps first-time vs. repeat visitors, heavy users vs. light users, etc.?
- What visitors expect from your website? Do men and women have the same expectations? Old and young? Do people who arrived as a result of a Google ad have the same expectations as those who arrived because of a TV ad?
- What channel your visitors prefer, and are there ways you can influence that preference so they frequent less costly, more profitable channels?
- How customers view your business, compared to the way non-customers view your business, relative to your competition?
- How your customer profiles and expectations change in response to market and broader economic conditions? And what, if anything, you need to change as a result?”
Colin Shaw in a recent post makes the same point in a different way. Here are some relevent extracts from his post:
“Google and Facebook, as Eli Pariser discusses as a part of TED, are engaged in the process of quantifying preferences from the timing and frequency of online clicks, and using this information to alter web content. The stated goal of this practice is to “personalize” the web experience.
A lesson gleaned from Dell’s 1990s laptop boom illustrates the point that preference and value are two different things. Dell let its customers customize all aspects of their computer’s hardware – from screen size to keyboards to RAM – everything but color. Nobody thought to customize color, because a laptop was supposed to be black or gray. However, when color laptops were introduced, sales skyrocketed and we all learned that color was indeed an important factor.
Imagine if Google and Facebook had monitored “clicks.” They would infer that because customers did not indicate a laptop color preference, it doesn’t drive value and is therefore irrelevant.”
Lesson 5: there is an assumption behind all predictive analytics and that is “all things being equal” – that is to say that predictive analytics assumes that the future will be a replay of the past.
Lesson 6: human behaviour is shaped by the ‘structures’ in which human beings are embedded, change the structure and you are likely to see human beings change their behaviour. Think about how the recession (e.g. job losses) have changed the shopping habits of consumers in the western economies.
Why won’t big data and analytics make you into the next Apple?
Apple was busy creating a new future (a break from the past) rather than exploiting the past. If you take a look at the US automotive industry the big US automakers were busy building and selling gas guzzlers because the analytics showed that these were the cars that Americans were buying. At the same time Toyota was busy living into a very different vision of the future: hybrid cars and electric cars. Who was right? According to the data and the analytics it was the US automakers. What would you say now? Toyota?
If you are not inventing the future you can still prosper by picking up the weak signals that point towards a new trend. I once asked Bob Greenberg (R/GA) the secret of his success and he told me it was his ability to see these trends and act upon them before others. You might imagine that analytics might help you to spot trends. My experience of traditional analytics is that the modelers do all they can to strip out the outliers and create a normal distribution so that the maths works – in doing that they filter out the ‘weak signals’ that point towards these trends.
Perhaps it is best to end by remembering what Colin Shaw points out: how would analytics have disclosed that customers wanted to customise the colour of their laptops and that once this option was made available then Dell’s laptop sales would surge.
What do you think? If I have it wrong then please do educate me.
Posted on October 24, 2011, in Case Studies, Customer Insight (inc VoC), Customer Strategy and tagged Analytics, Apple, Big Data, customer insight, data, Dell, Dyson, inventing the future, outliers. Bookmark the permalink. 3 Comments.