Using data to make better business decisions
Early-stage companies that understand the value of data analytics and develop a business model that intelligently uses insights gained from the analysis of their data often get to market faster, have better products, and are more efficient in their operations. As a company grows, it automatically collects a lot of interesting data about its market, its customers, its employees and its financials. Analyzing this data to discover interesting trends and patterns can be a great opportunity to improve every aspect of your business model. For many entrepreneurs, insights based on facts should trump everything else — gut instinct, best practices, job titles and MBA degrees. This article is a guide to incorporating data-driven decision-making into your startup.
As technology improves, we are able to collect, store, and search vast amounts of data about every esoteric aspect of our culture and economy. When you couple this data with smart algorithms that can analyze it to identify interesting patterns, novel business insights can emerge. Entrepreneurs are learning to utilize this capability to make better business decisions. Big data has a number of use cases in areas as diverse as product improvement, customer service, brand development, HR, accounting, and security. Acting on the insights generated by data analysis can save money, improve the product, or increase revenue through more effective targeting of customers.
What is Big Data?
More data has been created in the past few years than in the rest of human history combined. The scale of this data is such that it has to be measured in yottabytes, or chunks of 1024 bytes. In a world in which pervasive sensors embedded in cars, phones, wristbands and other devices are constantly creating new data, this trend is only set to accelerate. The Internet of Things will soon become the largest single source of data in the world. Much of this data is unstructured, meaning that it cannot be stored in traditional databases.
As the cost to collect and analyse this data becoming cheaper by the day, businesses have an unprecedented opportunity to develop information about what works, what doesn’t and why. Agile companies that adjust their business model and strategy in accordance with data are increasingly taking business away from those who disregard it. Whether a company is trying to become more efficient, optimise its workforce, improve product quality or drive revenue, big data can provide deeper insights than traditional systems because the new data is richer and more voluminous.
Data-driven decision-making (DDDM) is an increasingly popular management strategy that values decisions that can be supported by big data. “Data-driven” simply means that decisions are made on the basis of empirical evidence, as opposed to intuition or anecdotal experience. DDDM gives a business an advantage over less technical competitors – A study by the MIT Centre for Digital Business found that businesses engaging in DDDM were 4% more productive and 6% more profitable than competitors.
But Big Data is not a panacea. In practice, business decisions are never based solely on cold data, instead it is often one of the many components that ungird a firm’s business decisions. “Data-informed” is the term used to describe a governance strategy that takes empirical data into account, but does not disregard intuition and personal judgment altogether.
Being data-informed, as opposed to 100% data-driven, is a way of acknowledging that a business never has all the data at its fingertips to make a perfect decision. While decision-makers might have enough data to lead them towards a local maximum, the global optimal strategy for the largest market often requires a degree of human judgment, which is less quantifiable and more error-prone.
Use Cases of Big Data
Here are three possible use cases for big data applications in a business context.
Improved understanding of customers
With the rise of big data, there are more contact points between a business and its customers than ever before. Businesses need to analyse all of these data points to gain a fuller understanding of their customers; this in turn drives engagement, loyalty and revenue. Among other things, a business that engages in intelligent data analysis is better able to segment its customers, more likely to offer customers targeted cross-sells or up-sells, retain its most valuable customers, identify loss making customers and deliver superior overall service by analyzing the feedback from happy and unhappy customers.
Cyber security
As society has become more dependent on the Internet, so the threat posed to businesses by computer fraud, phishing scams, espionage and even cyber terrorism has increased. Big data technologies offer businesses and law enforcement a way to analyse social media, emails, VOIP and other telecommunications in order to predict threats, identify them early and respond in real-time. A customer’s interaction signature patterns can be used to identify fraudulent transactions or attacks on the firm’s infrastructure.
Operations analysis
The data produced by computers, sensors, meters and GPS devices inside a firm can be very useful to it in improving its operational efficiency. This data can be used to identify productive employees, reduce transportation costs, identify cost savings by changing the production schedules, weed out expensive or poor quality vendors, etc. This type of analysis enables businesses to improve understanding of customer behaviour, optimise customer service, and identify any anomalies or opportunities in day-to-day operations.
Advantages of Data-driven Decision-making
In each of these and other similar use cases, DDDM offers businesses a number of advantages over competitors that use traditional management strategies.
Reduced time to market
For instance, drug manufacturers use big data to simulate clinical trials, lowering costs, speeding up learning and reducing patient suffering during trials. This means that a drug life cycle that might have previously taken five years can be reduced to two, benefitting manufacturers and patients alike.
Improved Product
Netflix and Amazon are able to provide targeted recommendations to their customers by analyzing their prior purchases. Facebook and LinkedIn are able to identify the people you are most likely to want to be “friends” with. Google is able to predict the query you plan to type by just from the initial key strokes. Every one of these product features are enabled by Big Data.
Optimized workforce
HR departments use big data to better predict which job applicants will be a good fit for the company. Known as talent analytics, this helps HR reduce costs associated with employee turnover and determine what hires the company needs to make and when.
Improved financial management
Corporate finance departments use big data to identify risky customers, monitor suppliers, flag credit risks, and thwart fraud, based on trends in historical data. There is no better way to deal with business partners that don’t pay on time than to avoid starting a relationship with them in the first place because the data unearths a red flag.
Fewer equipment failures
The transport industry uses big data to predict when maintenance is required for planes, trains and cars. This is a way of avoiding unnecessary delays and negative customer feedback. Ideally, companies want to understand the cause of equipment failure and how it can be resolved. If they can achieve this before sending a maintenance professional, they will save money.
Case Study: A/B Testing
Once a firm has been alerted to a potential opportunity for improvement by data analytics, A/B testing is one good approach for testing a new hypothesis that the firm might have concerning how to optimize its service.
An A/B test begins when a company uses surveys, feedback boxes, usability tests or other data to determine 1) the existence of a problem and 2) an hypothesis on why the problem is occurring. For example, the vacuum cleaners sold in Singapore by a multinational company might be returned frequently. Based on the reasons given by the customers returning the product, the firm may determine that the root cause of problem lies in the plastic being used to produce the vacuum cleaners which cracks easily. At this point, company executives make a prediction: a change of plastic supplier would result in fewer products returned (and thereby higher profits). They might test this by selling products made using the old plastic alongside products made using plastic from a new supplier. If their analysis is correct, the products manufactured using the new supplier’s plastic should be returned less frequently.
A/B testing is often used in the context of a company’s website but can be used to validate any aspect of a firm’s business. The following is a list of things that can be easily subjected to an A/B test:Calls to action on a website
Calls to action on a website
Testimonials from customers
Headlines for press releases
Colour schemes for products or website
Sales funnels
Page layout
Pricing for products
Purchase discounts
Signup bonuses
Marketing collateral
Press release format
Companies should first focus on the proverbial “low hanging fruit” and find their big wins by focusing on the major problems they uncover through the analysis of customer feedback such as product defects, quality problems, etc. Although it is less likely to be a point of contention with customers than a faulty product would be, A/B testing of pricing is another area that can produce especially significant financial results. Once the big win areas have been optimized, a firm should keep these principles in mind to fine-tune the finer parts of their business. The key lesson is that the firm’s leaders should view data as a critical element to their decision making. They should adopt a mindset where every time they have a question whose answer is uncertain, their default instinct should be reach out for a data driven answer to it.
As regards A/B testing, it should be conducted on the basis of variables that have been uncovered as a result of data analysis, as opposed to variables chosen arbitrarily by management. To ensure maximum accuracy, keep the A test as close to the B test as possible, changing only the variable that you are testing. Run tests on a small portion of your market to prevent overspending and alienating a large proportion of potential customers or leads. Consider multiple rounds of tests if time and budget permit.
Conclusion
A data-driven approach to management is an essential tool for startups that want to become efficient and competitive. It has a wide variety of use cases and can bring benefits to departments as diverse as accounting, product development, customer service, security and logistics. Once armed with insights from data analysis, companies should formulate theories about how to optimise their service and use tools such as A/B test to validate these theories rigorously. However, a thoughtful manager should not make decisions based solely on empirical evidence; she should also supplement it with her personal knowledge, intuition and individual life experience.
