Chapter 5
In This Chapter
Understanding the data types available
Appreciating where to get data
Peeking into data’s future
The world is becoming increasingly datafied. Being datafied means that an activity or action (such as purchasing something) has been turned into data. This is also known as datafication.
Datafication is leading to new forms of data. Some of the data we can now collect is new, some has been around a while but we’ve only just found ways to make it genuinely useful. In this chapter I look at the main forms of data that are being collected, stored and analysed.
In this chapter, I give some examples on how you might use these types of data in your business. There’s more on big data uses for small businesses in Chapter 3. To understand the technology advances that underpin these new forms of data, see Chapter 6.
Of course, this information is extremely useful to booksellers, publishers and authors. As an author, I’d love to know if I was losing my audience at a particular part of my book! This information could prompt a second edition and hopefully a more committed readership. Publishers could figure out what is being consistently highlighted and establish whether those things indicate trends around which they could commission a new book. And booksellers could alert the reader when he’s getting close to the end of the book with recommendations about what to buy next based on his previous purchases. Amazon, of course, already does this.
Essentially, activity data is the computer record of human actions or activities that occur online or in the offline physical world.
Unless you’re a Luddite, completely disconnected from and opposed to technology, living in a shack on some remote island, paying for everything you need in cash (unlikely, considering you’re reading a book on big data!), you’re creating data in your wake. This data is valuable. Analysing it provides insights on everything from your personal health to crime prevention to increased sales, improved customer service and elevated business performance.
Think about your typical daily activities: Most of them leave some digital trace (data) that can be and is collected and analysed. If you make a phone call, a trace of that call exists as call data; depending on whom you call, the content might also be recorded. If you go shopping and use your credit or debit card, the trace is transaction data – what you buy, when, where and for how much. If you browse the Internet, a digital record is created showing where you accessed the Internet from (using the computer ISP, or Internet service provider), what sites you visited, how you navigated through them, what held your attention and how long you stayed. Even if you choose not to work and go for a run instead (pat on the back for you!), if you wear a Fitbit or UP band your movements create data showing how fast you moved, how many steps you took and how many calories you burned. CCTV (closed-circuit television) cameras also probably recorded you at various locations as you enjoyed your run!
Activity data allows you to monitor how your customers interact with your business, brand, products and services over time.
The main advantages of activity data are:
The main disadvantages of activity data are:
The applications for activity data are as endless as the activities about which data can be collected and analysed.
For example, Jawbone, the company that makes the activity band, now collects health, movement and sleep data from millions of people around the world (including me). This means that Jawbone has access to 60 years’ worth of sleep data – every night! No company has ever had so much sleep data, which means that Jawbone is able to analyse the data to understand more about sleep, sleeping patterns and what disrupts those patterns. Jawbone could, for instance, look at the data and work out how many hours of sleep are lost, on average, when the Super Bowl is broadcast in the US or how long it normally takes for business travellers to return to normal sleeping patterns after international travel.
How best to get your hands on activity data depends on what activity you seek to understand and whether that data currently exists or could exist in your business.
You could ask your target audience questions regarding the activity you want to understand more about. This can be done through a market research survey or focus group. Or you could put a mechanism in place to capture their answers as part of the activity in question. For example, if you want to know more about why someone bought a particular product online, you could add the question to the checkout page so he needs to answer it to complete the purchase process.
Conversations are increasingly leaving a digital record – either through the text you write in an email, blog, SMS (text) message or social media post, or as an audio recording of a telephone, teleconference or Skype call. This is known as conversation data.
Conversation data can be extremely useful for small business owners because it provides insights into how happy or otherwise your customers, clients and suppliers are.
The main advantages of conversation data for small business are:
The main disadvantages of conversation data for small business are:
Conversation data allows you to know what is being said about your business in conversations between your business and your customers, clients and suppliers.
This information can shed valuable light on shifting buying patterns and changing requirements, highlight service or performance issues in the business or indicate new product or service development ideas.
Sentiment analysis or opinion mining seeks to extract subjective opinion or sentiment from text or audio files. In any language, some words are neutral, others are considered positive and others still are more negative in tone and meaning. These variations in sentiment or meaning can be identified by text-based sentiment analytics. Audio-based sentiment analysis considers the pitch, tone and stress levels in a person’s voice to determine how he’s feeling and reacting to the conversation he’s currently involved in.
Beyond polarity sentiment analysis goes farther than ordinary sentiment analysis further by classifying the emotional state of the person writing the text, such as ‘frustrated’, ‘angry’, or ‘happy’. This type of analytics is becoming increasingly popular with the rise of social media, blogs and social networks where people share their thoughts and feelings about all sorts of things – including companies, brands, products and services.
Millions of people are leaving reviews, rating products, making recommendations and expressing their opinions about businesses, products and services. Your customers or potential customers are using those opinions to influence and direct their purchasing decisions. It follows, therefore, that you absolutely need to know what people are saying about you and sentiment analytics can help you do that.
Large companies like Gatorade and Dell recognise the importance of tracking conversation as a way to market their products, identify new opportunities and manage their reputations. As a result, they have invested heavily in social media command centres.
Wherever you are currently having conversations there is an opportunity to collect conversation data.
Clearly social media platforms are critical for conversation data, and they provide a wealth of information that can then be analysed for future improvements. A good way to start might be some simple sentiment analysis (perhaps using the software mentioned in this chapter) to find out what your customers are saying about your product or service online.
The volume of photo and video image data generated today is staggering. The explosion of content largely comes down to digital cameras and, more recently, smartphones coupled with our ever-increasing connectivity and our insatiable desire to share content on social media platforms.
In addition to all the photo and video data created and shared via smartphones, there is also all the CCTV camera footage. In days gone by, companies may have recorded their retail or storage premises for security purposes, but the recordings were never stored long term. The recording was made to tape and then after a week or so the tape would be used again and new recordings would be made over the old ones.
Now some of the more data-savvy stores are keeping all the CCTV camera footage and analysing it to study how people walk through the shops, where they stop, what they look at and for how long so they can make alterations to offers and boost sales. Some are even using face recognition software, so it probably won’t be long before a combination of data sources such as CCTV camera footage, loyalty card information and face recognition software will have you being welcomed to a store on your smartphone and directed to particular special offers or promotions that may be of interest based on your previous buying habits!
The main advantages of photo and video image data for small business are:
The main disadvantages of photo and video image data for small business are:
The applications for photo and video image data are endless and they already impact crime prevention and health – as well as business.
If your business doesn’t have retail premises, then the applications for video data may be limited for you. Photos can provide useful information for many different types of businesses – photo metadata, for example, tells you who took the picture and when and where it was taken. But brands can now use photos and video data to detect where it is used. For example, a clothing brand can start searching for photos that show someone wearing its brand logo or restaurants can search for photographs taken inside or outside their premises. This allows them to identify and potentially engage those individuals.
By pinpointing where the data was coming from in the Philippines (using GPS sensors in the photos or through the text) the work of the volunteers then creates an online map, not just of the disaster zone but of the needs in each area.
The result was that when the disaster relief effort arrived in the Philippines they didn’t need to waste days working out what was happening and where the worst hit areas were. From the map created by people halfway around the world, relief providers already know who needed water, who needed food, where the dead bodies were, where people had been displaced, where the most damage was and what hospitals were least damaged and therefore more able to help the injured.
Video and photo data can be obtained by simply starting to collect it using digital cameras.
If more advanced analytics are required, it often means installing purpose-built systems. For example, if a small shop already has a network of CCTV cameras installed then the data can be brought together for an analysis of how customers walk through the shop, where they stop and which parts they avoid. Testing the existing data shows any gaps where new cameras or systems need to be added to improve the analysis.
Photos can often be collected and analysed from the vast amount of photos shared on social media sites such as Twitter, Facebook, Pinterest and many others.
A vast amount of data is generated and transmitted from sensors that are increasingly being built into products. This is known as sensor data. For example, your smartphone is smart because of the inclusion of various sensors that capture data. Your smartphone contains:
There are also sensors in the natural environment. For example, there are sensors in the oceans for measuring the health, temperature and changes in the oceans in real time. In Japan, there are sensors in the soil to collect data on how healthy the soil is. Companies are combining that data with weather data; farmers can then subscribe to the service to get information to optimise yield, including when to put fertiliser on the crop and how much.
Sensor data can be extremely useful for small business owners because it can help to answer some important strategic questions. For information on how to identify your key strategic questions, see Chapter 11.
The main advantages of sensor data for small businesses are:
The main disadvantages of sensor data for small business are:
Not only did this fashion retailer massively increase sales by getting smart about its data requirements and combing small traditional data with untraditional big data, it used the insights to make a significant saving by closing one of the stores. The sensors were able to tell them the store owner that the foot traffic reported by the market research company prior to opening in that location was wrong and the passing traffic was insufficient to justify keeping the store open.
Increasingly, more and more machines are equipped with sensors to monitor performance and provide information on when best to service or repair the machines.
For example, modern cars are full of sensors that measure everything from fuel consumption to engine performance. These sensors allow for dynamic servicing and better long-term performance. On-board sensors also alert the driver if he gets too close to another car or object and can even parallel park the car without the driver doing anything!
In retail, data has long been collected via bar code – however, the sensors known as radio frequency identification (RFID) systems increasingly used by retailers and others are generating 100 to 1,000 times more data than the conventional bar code system.
All the measurements are stored in the computer and streamed via satellite back to Rolls-Royce headquarters (HQ) in Derby, England. And that’s true for the entire fleet of Rolls-Royce engines – which is a lot of data when you consider that a Rolls-Royce-powered engine takes off or lands somewhere in the world every two and a half seconds.
Whenever those thousands of engines are in the air, they’re gathering data that’s constantly sent back to HQ and constantly monitored using clever data analytics that look for anything unusual occurring in the engine or any sign that it may need to be serviced early or repaired. In Derby, computers then sift through the data to look for anomalies. If any are found, they’re immediately flagged and a human being checks the results and, if necessary, telephones the airline and works out what needs to be done – often before the issue escalates into an actual problem.
These sensors therefore allow for dynamic maintenance based on actual engine-by-engine performance rather than some automatic rota system based on time alone. Instead of pulling an expensive piece of equipment out of service every three or six months, these sensors allow the airlines to maintain their fleet much more cost effectively, and, more importantly, these sensors make the planes much safer.
Sensor data may be accessed either by using sensors inbuilt in devices you already have or by installing new sensors.
The world is filling up with sensors. Smartphones can contain GPS sensors, accelerometer sensors, light sensors, fingerprint readers and so on, all readily used by any small company. Smart watches contain all these features as well as heart-rate sensors. Soon, most devices, from diapers to clothing, sports equipment to cars and thermostats, will generate sensor data you can use for business purposes.
The Internet of Things (IoT) is coming about as a direct result of more objects being manufactured with embedded sensors and perhaps more importantly the ability of those objects to communicate with each other.
International Data Corporation describes the Internet of Things as ‘a network connecting – either wired or wireless – devices (things) that are characterised by automatic provisioning, management and monitoring. It is innately analytical and integrated, and includes not just intelligent systems and devices, but connectivity enablement, platforms for device, network and application enablement, analytics and social business and applications and vertical industry solutions. It is more than traditional machine-to-machine communication. Indeed, it is more than the traditional Information and Communications Technology (ICT) industry itself.’
Essentially, the Internet of Things explores what is and will be possible as a result of advances in smart, sensor-based technology and massive advances in connectivity between devices, systems and services that go way beyond business as usual. For example, according to estimates from research groups such as Gartner and ABI Research, by 2020 there will be between 26 and 30 billion devices wirelessly connected to the IoT. And the resulting information networks promise to create new business models, improve business processes and performance while also reducing cost and, potentially, risk.
The day will come, not far from now, when your alarm will be synced to your email account and if an early meeting is cancelled, your alarm will automatically reset to a later time which will also postpone the coffee machine to the new wake-up time. Your fridge will know what’s in it and place online orders to replenish stocks without you having to do anything. When you come home from work, your fridge will tell you what you can make for dinner based on what you currently have in stock.
The wired and wireless networks that connect the Internet of Things often use the same Internet Protocol (IP) that connects the Internet – hence the name. These vast networks create huge volumes of data that’s then available for analysis. When objects use sensors to sense the environment and communicate with each other, they become tools for understanding complexity and responding to it quickly. The resulting physical information systems are now beginning to be deployed, and some of them operate without human intervention.
Ever-smaller silicon chips are gaining new capabilities, while costs are falling. Massive increases in storage and computing power, some of it available via cloud computing, make number crunching possible on a very large scale and at declining cost.
18.220.181.186