Post 22 May

Using Data Analytics to Make the Leap From a $10 Million to $50 Million Company

In essence:
Data is the most important resource in the world economy, and it may help small firms develop significantly.

Organizations can collect data about their operations and evaluate it to improve decision-making by making minor changes to the existing business processes. Business executives assess the findings after the procedure is complete using good human judgement.

Consideration of the issues that are most crucial to the organisation should be the first step in any effective data analytics programme. Rigidity in gathering data pertinent to those queries and ongoing improvement through iteration are other must-haves.

The Economist’s May 2017 cover included oil rigs with names like Google, Facebook, and Amazon that resembled office buildings. The article’s bold assertion is that “info, not oil, is now the most valuable resource in the planet.” The oil-data parallel has long been a favourite in the corridors of Wall Street, Silicon Valley, and Washington.

To continue the comparison, data developed the same value as oil. Oil is useless on its own. It doesn’t offer anything to eat, to stay in or to enjoy. Yet as mankind discovered how to extract, refine, and utilise its energy, it changed the globe. Because of human inventiveness, oil has value. Data is the same way. Massive data sets are impossible to manage and less valuable than anecdotal client conversations for an unaided businessperson. But, data offers enormous value when used in conjunction with the appropriate analytical methodology and analytics technology.

Almost any information and insights that are crucial to a business can be discovered through proper data gathering and analysis. It can provide information on your clients and staff, business sector and operations, history, present, and — occasionally — future. Data can be crucial in helping developing firms advance to the next stage. The data analytics equivalent of an oil rig, a refinery, and a combustion engine must be built in order to fully utilise that power.

The good news is that oil rigs are simple to construct in the domain of small business data analytics. Some businesses even create them by mistake. Businesses tend to utilise more software as they expand. Software like this creates “data exhaust,” or digital recordings of business actions that you can arrange and analyse. This software ranges from simple accountancy to enterprise resource planning ERP systems. With a plan, data analytics becomes even more effective: You can collect data in a way that addresses your most pressing concerns and makes it simple and quick for you to conduct your most crucial analyses by making a small adjustment to the information technology that most companies already employ. From then, businesses can go above and beyond and gather further data through instruments like surveys and publically accessible sources, or they can buy pertinent datasets.

In this chapter of the Project $50 Million series, experts offer guidance on how to use data to drive growth, including what questions to ask, how to collect data, how to analyse it, and how much to trust the outcomes.

An Opportunity for Data for a Developing Company
Launching a data analytics practise in a developing company can, among other things,

Establish a positive cycle
As a company expands, more data is produced, more data can be collected, and more resources are available to mine that data. Also, when an organisation expands, the value of a tiny data analytics contribution increases.

Accelerate growth
Data may save businesses months of trial and error while providing insights that might not have been possible without it. But, firms that are new to data analytics may find it simple to ignore the right patterns or draw incorrect inferences. Or, to put it another way, believing a false data correlation or an incorrect interpretation of significance may be worse than performing no data analysis at all.

Find out more about your clients
Every savvy company wants to learn more about its clients. Customer interactions often yield the most insightful data for new enterprises. But as your organisation expands from $10 million to $50 million, consumer dialogues with business leaders won’t be able to keep up. But as the clientele grows, thoughtful data collecting and analytics can keep up. For instance, if you’re a retailer, you might solicit feedback from customers or gather information on those who have created online accounts. Several businesses use random customer sampling for surveys. Companies with direct client contact may have information in customer relationship management CRM systems, and with little training, their workers can learn to gather that data in a way that allows for later analysis.

Increase cooperation and trust
Not everyone will know, trust, or work well together as a company expands from a small core staff to a massive corporation. Data collection on personnel and operations can help managers better coordinate teams and gain insights into how your business is managing expansion. With effective analytics, some of that can happen automatically: As long as those measurements are well-chosen, transparency around them will focus their efforts and provide immediate, unbiased feedback if different company units use those metrics or KPIs as barometers for success. Employees, who as the business expands have less opportunities to contact with top leadership, may benefit from this clarity in objectives and data-driven decision-making.

Creating Business Value from Data
The first step in creating a data analytics process that will guide you in the correct direction is to identify the objectives of your programme and the part you want it to play in your strategy. Despite how simple that may seem, far too many businesses begin by using whatever data they have to run regressions on until insights emerge. Indeed, a corporation can quickly mine its existing data for insightful information. Yet over time, there is a rapidly increasing risk of going in the wrong direction by employing data and models that don’t correspond to the questions that are most crucial to your company.

8 Steps to Building a Data Analytics Program
1. Hire The Right Person Or People Who Know Statistics, As Well As How To Interpret Statistics And When They Can Go Wrong.
2. Determine Which Questions Matter Most To The Success Of Your Business.
3. Identify The Data That Is Most Relevant To Those Questions.
4. Collect Or Buy The Required Data.
5. Let The Statistics MavenS Analyze That Data Until They Find Insights.
6. Automatically Generate The Most Useful Metrics, And Distribute Them To All Relevant Stakeholders.
7. Iterate Steps 2 Through 6 To Refine Your Analysis And Dig For More Insights.
8. Accept That This Process Isn’t Always Linear. Sometimes A Later Step Informs A Prior One And You Must Iterate Again.
Determining those most crucial questions is thus the first step in developing the ideal data analytics approach for each firm. The next stage is to identify the data and analytical techniques needed to provide the organisation with the answers to those queries. There are numerous ways to learn about customers, for instance: A $10 million firm selling candy has a much different consumer interactions than a $10 million company providing software to corporations, both in terms of quantity and type. But, the first step in creating an analytics strategy that will be effective in the long run is to be clear about the questions you want to answer and the data sources you have or can generate to address those concerns. By understanding more about the seasonality of its sales, the candy company, for instance, may be able to respond to inquiries about how to best plan for Halloween and Valentine’s Day. Internally, it would begin by compiling and setting up automatic gathering of useful formats for sales data. The software corporation must search outside of itself for information if it has concerns about a new target consumer for the upcoming release of its product. A mix of focus groups, polls, third-party data sources, social listening, and other methods may be used by the leaders there. These inquiries from new customers won’t be answered by data from the business’s previous and present operations.

Seven Questions That Data Analytics Can Address
Here are the seven most important concerns that small, developing businesses should consider, customise for their unique requirements, and then configure their data analytics procedures to address. Each begins with three typical customer-related inquiries and contains thought processes to assist you in moving from question identification to data analysis:

1. Does our company draw in the correct kinds of customers?
Effective data analytics can identify discrepancies between actual consumers and the ones you anticipate, thereby assisting you in avoiding pitfalls or taking advantage of unforeseen opportunities. It can be challenging to translate findings into insights. Consider the case where one of your product’s target consumers is college grads beginning their first jobs. Data collecting indicates that you have the correct target market, however the majority of your consumers are men. What’s next? Are there any aspects of your product that appeal to males rather than women? So, a more targeted approach to guys in advertising may be the best course of action. On the other hand, perhaps your messaging attracts to men more than women and you need to make your advertising more diverse. Or maybe your company isn’t at all to blame for the gender gap; perhaps a rival product is actively courting female consumers. Your data analytics can identify the problem, but in order to formulate a suitable answer, you will need to conduct additional study.

2. What is the customer pipeline yield?
If your salespeople track each lead from initial outreach through purchase decision, a robust analytics software can determine the success of your sales efforts at particular pipeline stages. You can update statistics like: X% of leads result in actual sales discussions, Y% of leads result in trials, and Z% of leads result in paying clients. If X starts to decline, you might not be focusing on the proper clients. But if Y begins to decline, perhaps your sales team requires additional training. Your main value proposition can be flawed if Z starts to fall and you haven’t modified the way your trials operate. This incredibly useful knowledge can be obtained with little effort. If anything, it typically just entails minimal standardisations and adjustments to the way your sales staff monitors their progress. For instance, it takes little effort today but will provide many answers tomorrow to preserve extremely brief records of when and why a sales team ceased chasing a potential customer.

3. From what sources and why are we losing clients?
Some businesses take every step possible to get thorough data in this area, including conducting exit interviews, follow-up surveys, looking into service failures, and more. Even though it would be impractical to look into every exit if you had thousands of customers, it is almost always worthwhile to look into the reasons why people leave. Do people typically depart when your pricing go up? During a certain season? in response to the introduction of a rival’s product? in response to the introduction of your product ex. New Coke? Some pertinent data can be gathered automatically, such as the timing of subscription cancellations, but other data may need to be gathered manually through interviews or questionnaires. If you have any questions, you can achieve this goal by asking them during the cancelling process or by sending them an email. Retail items for which customers must opt in to repurchase rather than out make it more difficult to gather this data. Selling bedsheets, for instance, makes it very challenging to predict when customers would transfer to a different brand; nevertheless, selling a monthly membership to a magazine makes this task much simpler. Yet regardless of the situation, it’s virtually always worthwhile to take the time to learn why clients depart.

4. Do we make competitive offers and hire the right people?
It can be difficult to choose the appropriate candidates, but there are signs that something might not be right. For instance, high turnover can be an indication of a number of problems, but over the long run, the rate at which you dismiss employees and/or mutually separate is almost certainly a sign of the efficacy of your hiring process. You may tell whether your hiring programme is getting better, becoming worse, or staying the same by looking at retention and performance by cohort, or groups of people hired relatively close together. This becomes increasingly crucial as businesses expand. Because you need to go for outside data, comparing your offers and salaries to those of competitors requires more work. Yet, it may be worthwhile, particularly during periods of rapid expansion. Spending more time on challenging data collection when you’re already short-staffed may seem contradictory, but if non-competitive compensation results in a low-yield recruitment season, you’ll wish you had assigned your best workers to the assignment.

5. How might different scenarios affect our financial results?
It’s a little trickier to respond to this issue than to generate a statistic from streams of data that already exist. Yet, performing “what if” assessments gets simpler once a corporation uses software to track and manage a sufficient number of its operations. In some circumstances, it’s as simple as entering a few figures to see how your spreadsheet, accounting programme, or ERP system responds. Data analytics can help you estimate the links between moving pieces in complex scenarios, which may demand a more thorough understanding of how they interact. An easy scenario to assess is “What happens to our bottom line if sales decline by 20%?”. More complex statistical models and assumptions are needed to answer questions like “What will happen if China imposes a tax on our industry,” as well as likely external data.

6. How can we make the most of opportunities for growth?
One of the most common and significant uses of data analytics is to provide answers to this large group of queries. Data analytics can be used by businesses to quantify the virality of a new innovation, determine how much surplus capacity they can allocate to promotional freebies, or A/B test referral incentives. You can take a data-informed approach to exploring new geographical markets, introducing new product lines, and even improving the way you invest in potentially game-changing innovations by casting a wider net for data collection, such as by conducting a survey, purchasing external data, or utilising free government data. You might need to conduct some one-time data collection to answer these more comprehensive inquiries, looking for data that you don’t intend to routinely obtain for tracking purposes. This is particularly typical when a new geographic market is being entered. But, once you’ve made the decision to join a new market, you won’t need to continually reevaluate your initial assessment of whether doing so may be advantageous for you because you’ll be gathering information on your actual experiences and results.

7. What blind spots do we have? What more sorts of data do we require?
It’s an almost universal truth that as you proceed with data analytics, you’ll discover needs that weren’t anticipated. You’ll ultimately find yourself wanting to answer questions that your data doesn’t speak to but that you could answer with a little more work, whether it’s new scenarios forming or natural follow-ups to the analytics you’re already running.

This can often lead to new lines of investigation, such as when a competitor is focusing on a certain demographic and you need to determine why their offering or messaging is drawing away your clients. But, blind spots might occasionally be present as important nuances within the questions you are already asking. Several businesses provide infant parents with things to make their lives simpler and safer, yet some people feel that the term “parents of infants” is too general. While certain goods may be more appealing to parents juggling an infant and older children, others may be better suited to “new parents” who have just had their first kid. Businesses that target “parents of newborns” broadly may waste a lot of time and money talking with parents who aren’t really part of their target market. The correct inquiries and data gathering can alter how a business views its clients and their issues, which will alter its return on marketing expenditure. Even if not every item on the wish list will be worthwhile to pursue and not every blind spot may be revealed, maintaining humility, openness, and curiosity will help you make continuous advancements in your data analytics.

Construction of the Data Analytics Your Business Requires
It’s time to create the data equivalent of your company’s internal combustion engine, the gear that transforms data into insights and propels you in the direction you want to go, now that you’ve committed to a data-informed approach and considered the essential questions. Below are the most crucial instructions.

Employ team members with a statistical background.
Make sure the data is telling you what you think it is telling you if you’re going to rely on data analytics to make critical business decisions. You need someone who is knowledgeable about running the analytics as well as how to evaluate the findings and put them in perspective, including what you can and cannot infer from the data. You need someone who can, ideally, convey the implications to various stakeholders while being mindful of the methodologies’ assumptions and limits. Midsized firms frequently struggle with how to fill these positions; some believe that a candidate’s aptitude for math or technology automatically qualifies them for data analytics, but those are two whole separate skill sets. You need someone with experience in data science, social science, analytics, statistics, or a related field. Yet, the scientifically minded specialists in those categories are explicitly trained to consider how human factors and choices effect the dependability of data analysis and how to overcome such obstacles. Companies may not use social scientists very often, with the exception of economists.

Some businesses prefer to begin by cultivating analytics talent inside. They frequently start with a communicator with potential talent and interest in these skills and give them the training, tools, and courses they need to become analysts. It is ideal for analytics applications that must handle data, compute KPIs, and offer feedback and trend analysis. But if the issues get trickier, a corporation might require team members with more formal education. When making business decisions that demand a high degree of certainty, questions of causal inference—where you don’t only need to know whether X and Y move together, but whether X causes Y to move and to what extent—require additional skill. Businesses that don’t require a full-time employee for this employ consultants or, more efficiently and economically, a PhD candidate who has the necessary skills but hasn’t finished their dissertation.

The jobs within data analytics teams may specialise as they expand. In certain businesses, personnel with expertise in data management and collecting collaborate with employees with expertise in data use, in addition to frequent contractors or consultants who are called upon as needed. The ideal employee and, eventually, the ideal team, will vary slightly depending on the business. Yet one thing is always true: If you start by hiring good people, the remainder of the process will run much more smoothly and dependably.

Use an agile methodology.
Sometimes this entails actually putting Agile into practise, a methodology that is popular in software development and has been appropriated and used in corporate operations. Agile methodsopens in new tab centre an iterative approach to problem-solving for circumstances when you may not have a clear understanding of the solution with structure and discipline. Plan-do-check-act PDCA cycles, which are covered in more detail in the following section, are comparable to agile iteration. Agile can be used literally or not, but it always requires disciplined flexibility and an iterative, scientific methodology.

Reevaluate your KPIs.
An effective data analytics engine investigates the KPIs that most significantly shed light on the issues covered in the preceding section. The question debate demonstrates that choosing the KPIs that matter the most to your company is a unique process. Don’t be complacent after you have those KPIs, either.

Rethink your KPIs in the following three ways on a regular basis:

Go beyond the conventional measures.

This is not meant to belittle the widely used measures. A few common measures, such customer lifetime value, churn rate, and customer acquisition cost, for instance, may tell you practically everything you need to know about the financial viability of a software-as-a-service company. The Net Promoter Score NPS is a tried-and-true metric that may be used by businesses across all industries. Yet, there are also particulars regarding your operations and difficulties with your scenario that are specific to your company. Finding out whether odd or even unique KPIs make sense for your company, similar to how passenger revenue per available seat mile PRASM makes sense for an airline, requires effort and testing.

Think about how appropriate your KPIs are.
KPIs that worked well for you in the past might not function as well moving forward as your firm evolves and expands. Consider this: Are there crucial metrics that aren’t being recorded? Do we still use outdated KPIs that barely affect our bottom line?

Check to see how well your KPIs are doing.
Recognize how much you can rely on your numbers. Start with fundamental measuring attributes like:

Bias: Does a KPI routinely overestimate or underestimate what it is measuring?

How noise-prone are your KPIs? To put it another way, how long must you witness a change or how significant must the change be for you to be certain that it is significant? Sensitivity: If a KPI isn’t sensitive enough, you can’t rely on it to quickly reflect a significant change. Yet, a KPI that is too sensitive will sound the alarm on trivial events. Validity: To what extent does the KPI accurately reflect your actual priorities? For instance, innovation is notoriously difficult to measure with KPIs. Some businesses track patents or patent applications, but if you can hit that KPI, you’ve merely motivated staff to produce many little, patentable ideas instead of a select few significant and valuable ones.

Prioritizing KPIs merely because they are the simplest to measure is a typical mistake. Keep in mind that staff will react to these KPIs, especially if you are successful in creating incentives that revolve around them. The indicators that receive the greatest focus should be those that matter to you the most, and they should accurately and effectively reflect and summarise reality.

Consider a data warehouse instead of a spreadsheet.
Be prepared to manage data as a much larger organisation if you intend to grow rapidly. For a startup, one intelligent person managing analytics on a spreadsheet part-time can be effective, but not after businesses reach tens of millions of dollars. At that point, you would require teams of spreadsheet experts, and even then, there would be significant dangers to the usability, portability, and interoperability of your data.

It will be more difficult and expensive to make the changeover the longer you wait to deploy a long-term data storage solution, ideally the tidy, easily accessible, functionally formatted, central repository known as a data warehouse. Please don’t worry, spreadsheet aficionados. Excel data extracts are still exported by a decent system.

Make data governance official.
The formalisation of your data governance is another action item where delays can be expensive, much as the transition from spreadsheets to data warehouses. The teams that contribute to and use your data sources will expand along with the business. When you get to the point where the people who need to coordinate aren’t always working together directly, the clock is ticking. You’ll be better off if you establish formal, standardised procedures as soon as possible. Entrepreneurs often associate the phrase “data governance” with bureaucracy, but at its most basic level, it simply refers to the policies and procedures designed to ensure effective administration of an organization’s data assets. Strong data governance opens in new tab helps businesses extract the most value from their data, reduces risks related to the data, and ensures the quality of their data.

Add data and analytics applications piecemeal.
Growing businesses eventually become more able to afford and benefit from enhanced data collecting and analytics. They can support your growth by providing information about prospective new clients, target markets, and competitive environments. The secret is to add resources in a way that is manageable. Whilst the phrase “don’t bite off more than you can chew” may be overused, it’s wise in this case. Businesses can spend a lot of money buying resources that, as a result of trying to consume too much too rapidly, they either don’t use or use very ineffectively.

For instance, some businesses spend a lot of money on software they only use the essential features that are available elsewhere at no cost. Some, meantime, purchase large amounts of data from outside sources to “cover their bases,” even though the analytics team lacks the time or a clear application for it. Target your inquiries and the data you’ll need to respond to them with precision. Your performance will improve as you steadily and realistically expand your capabilities.

PDCA Cycle and data analytics
Several firms use the PDCA cycle, a four-step paradigm, to make business decisions and manage change. PDCA is short for:

Plan: Identify a difficulty or opportunity and come up with a solution.

Do: Try your strategy and get as much information as you can.

Check: Evaluate and summarise the testing and learning results, paying particular attention to how closely they adhered to the forecasts made during the planning stage.

Act: If your learning indicates a recommended action, take the advised action based on what you learned. You can repeat the “plan, do, check” cycle until an activity is obvious if things are still unclear.

Entrepreneurs who are familiar with Lean startup techniques may notice that its build-measure-learn cycle has many similarities opens in new tab. In fact, they are pretty similar. Both frameworks come from the scientific process, i.e., testing hypotheses and iterating until you acquire some useful learning, while as Lean startup language defines the concept for the unique use case of generating new products and services.

Understanding how data may enhance PDCA cycles and related frameworks is crucial for business decision-making in all areas. For instance:

Plan: Data analytics may assist you in identifying opportunities and even in projecting the outcomes of certain actions. You can identify planning goals and draw lessons from similar circumstances if you have enough data.

Example: A food company notices that cheeses constitute the majority of its top-selling items. As it digs deeper, it finds that a disproportionate fraction of its few vegan products—one in twenty of its offers, but six of the 50 fastest-growing—are vegan. The business plans to increase its marketing and vegan offers.

Do: You have a strategy in place, as well as some desired results. Now you must gather the relevant information in a method that will enable you to make deductions regarding the hypotheses included in your strategy. You can choose the type of data you need to collect if you know which analyses you’ll perform in the following stage to test the hypotheses you made in the previous step.

One assumption the food firm is aware of making is that consumers desire more and different vegan items. The business runs focused advertising campaigns for vegan items that are already on the market as well as vegan ones that are still in the works. Customers are invited to sign up for an email list to receive launch notifications on the landing page for the nonexistent products, and it also poses a few questions about what other products they might be interested in seeing.

Verify: Data analysis is frequently a significant part of this stage, and in other circumstances, it constitutes the entire process. Focusing your analysis on the issues, theories, and presumptions from your planning phase requires approach and dedication. Make sure the experiments you conducted and the questions you are asking are in line.

The food company, for instance, examines the experiment’s findings by examining the clickthrough rates on each advertisement and the feedback received from customers. That attests to the demand for additional plant-based solutions among consumers. The team was shocked to discover that many of the respondents to their brief survey weren’t vegans.

Act: At this point, the majority of people believe their data analysis work is complete. Time to put this into action. But keep in mind that you’ll eventually need to make more choices, put more changes into place, and do more PDCA cycles. Do yourself a favour and only declare a task completed once you’ve collected data that will guide your subsequent efforts. Finding out how effectively your testing predicted the outcome at scale is the initial step in this data collection; ultimately, your data collection will drive the “plan” step in your subsequent PDCA cycle.

For instance, now that its theories are largely confirmed, the company can invest R&D funds into new vegan products with greater assurance. However, it will make one crucial change: It will label the products as “100% plant-based” to appeal to non-vegan customers who are attempting to eat healthier or more sustainably.

How Much Should Strategy Be Based on Analytics?
This is a typical query, and you might anticipate an article on data analytics to provide “a lot” of answers. The finest businesses and businesspeople, however, are more data-informed than data-driven. They don’t always treat data and analytics as the only or even the most essential component when making decisions, but they do use them as significant inputs.

Data may provide all the information you require in some circumstances. Data analysis can speed up a business’s timeline and enhance performance if it is growing and attempting to accomplish more of the same thing for more clients while also doing it more effectively. But, there will be instances during the course of most enterprises when it won’t even be obvious which inferences teams can make from the information at hand. Yet there will be variables in some business decisions that you simply cannot capture with any form of data collecting, particularly the most critical and daring ones where organisations venture into unexplored terrain. Making future-related decisions requires taking a chance on the unknowable. The value of good data analysis in making those bets is enormous, but following its lead always and without question is to believe that every important factor can be captured in the data and fully comprehended by analytics.

Although data may be the most valuable resource on the planet, its full potential cannot be realised without human creativity, ingenuity, and intelligence. Data analytics may assist in moving your company to the next level and beyond if the correct team is put in place, your priorities are clear, and you commit to making decisions based on the data.