Digital Transformation for Small Businesses and Startups – Part One

The commonly accepted definition: “Digital Transformation is the adoption of digital technology to transform services or businesses, through replacing non-digital or manual processes with digital processes or replacing older digital technology with newer digital technology.”

Digital Transformation has a lot to do with the Internet of Things. It is fueled heavily by a massive and steadily growing number and variety of devices that didn’t exist a decade ago. Back in 2010, there were just 1.2 billion devices. According to Juniper Research, the world had some 35 billion IoT devices at the end of 2020 and is on its way to adding 48 billion more by the end of 2024. These IoT devices are literally almost everywhere and in everything.

We can divide IoT devices into about a dozen broad segments, in order of relative market size as:

1. Smart Industry
2. Smart Retail (to include eCommerce)
3. Smart Education
4. Connected Health
5. FinTech and Smart Services
6. Smart Vehicles
7. Smart Homes
8. IoT Wearables
9. Smart City
10. Smart Grid/Energy
11. Smart Agriculture
12. Smart Supply Chain

Digital Transformation is also closely tied to Big Data (volume, velocity, variety), and together impact our work processes and the skills everyone needs to effectively take advantage of it. Basically, if you’re in business, you’re in competition with others using IoT devices and technologies.

Deloitte conducted a survey in 2019 on the “digital maturity” of 1,200 companies with 500 or more personnel and annual revenues of at least $250 million. Their use of digital maturity was based on 7 points as defined below – according to their high-medium-low self-assessments. These were then collated into three tiers of maturity – 25% with high, 54% with medium, and 21% with low digital maturity. Companies with high maturity had net profit margins and growth rough three times higher than low maturity companies (45% – 15%). Medium maturity companies also outperformed low maturity ones by a factor of two to 1 (31% – 15%).

1. Elastic and secure infrastructure
2. Data mastery
3. Digital talent
4. Engagement with industry ecosystem
5. Smart workflows
6. Unified customer experience
7. Business model adaptability

It’s also worthwhile to take a quick look at how digital transformation is improving net revenue:

A few things can be said upfront. First, tech adoption isn’t and shouldn’t be linear – you don’t need to buy iPhones 1-10 in order to get an iPhone 11. You can leapfrog technologies. So, it doesn’t matter a whole lot where you’re starting. With a cogent strategy, you can catapult your company to be on the cutting edge of Data and Automation and achieve a high level of digital maturity reasonably fast.

The biggest problem, and why many digital transformations fail to deliver on their promises is that companies fail weighs on 3 main reasons:

1. Lack of a long-term plan.
2. Attempts to accomplish too much at once – disrupting their business in the process.
3. Failure to perform a due diligence evaluation and to prioritize improvements by, cost, advantage, and business impact.

Before you do anything else, you want to start quantifying everything you can about your business, whatever it is. You need metrics to compare everything against so you can prioritize your efforts to maximize your ROI (and to reinvest that ROI into further improvements to generate further improvements and ROI). This is important because digital transformation can be (though is not always) an expensive proposition. Some of today’s “Data as a Service” offers can literally save you millions of dollars and years of DIY effort to even come close to what they can provide.

If you measure twice and cut once, you’ll just about guarantee an ROI either from automation and/or data enhancements. Every change needs to be validated before implemented. I present this for the case of a distribution center, just from personal experience, though most points can apply to any physical production environment from factories to farms. This also relates only to the automation components, but the data elements afterward. When it comes to factory/production/distribution center automation, a good bit of physical engineering effort is involved to answer questions like:

• How long does it take to accomplish x-step in y-work process?
• What is your fully-loaded cost for paying someone to complete that work process?
• How much space in your facility does that task use? You need to make sure that any equipment you put there will actually fit.
• Does the space you are looking to connect already have a power outlet? If it doesn’t, you’ll incur additional costs (and possibly disruption), getting a suitable line to it… unless you want to daisy chain power outlet extensions (and create safety hazards). No, that ever happens!
• Does that space already have access to an internet/intranet connection – and what type? Like Bluetooth, Ethernet, Wi-Fi, LPWAN, Cellular, Satellite – Smart Agri IoT devices, for example, can be quite remote.
• To what extent do the upstream work processes impact the task? If you automate a wrapping station, optimal throughput will depend on always having enough packages to wrap.
• To what extent will the automation impact downstream work processes? Doubling your wrapping capacity may not be too productive without also doubling your packing output unless it’s already faster.
• Helpful to examine if the new automation will create a new job title, whether it will require training or a new hire. Take care to examine if there are other positions to which you can move anyone displaced by automation.

From experience with an Amazon DC in 2004, managers tried to maintain a certain ratio of workers for each position, across all work processes, from receiving to stocking, picking, QC, sorting, packing, and shipping, along with other support personnel. Typically, automating a given workstation can provide as much as a 400% increase in throughput over manual labor. That’s a machine that you only need to buy once, with some ongoing maintenance and energy costs.

1. Create a scaled schematic to make sure that everything fits,
2. Verify that the system requirements of each machine in the chain will fit/work together.
3. Compare your available options – for all of the stations you intend to automate. You’ll probably have several options, each with their own features, and some available on an a la carte basis, with others being part of an entire line.
4. Conduct a simulation – a live walkthrough of the process to include the station and at least one task up and downstream from it. Quantify the difference in time and money.
5. Assess the difficulty and expense (labor, amount of disruption, downstream impact, time to swap) to install the automated hardware. Also examine dependencies that by automating one station inherently requires another to be automated.
6. Set up a long-term budget and plan to implement your automation plan – start on training personnel, and establishing production targets and KPI’s to evaluate project success.

Big data is a huge component of digital transformation, too, but the burden of effort shifts to software engineers and data scientists. When it comes to implementing big data for your company, you need to ask questions like:

• How is your company accessing data now – or how do you store it?
• How many different sources of data are you using (and for what data)? If your business is in the Consumer Packaged Goods market, you could have a few hundred different data points. If you’re also into Revenue Management, you’re also tracking a lot of data about your competition and their products, possibly on a per channel/retailer/country basis.
• How much data do you receive daily? Could be just a few Mb… but could be Gb, possibly TB.
• What format and file types does your data come in (some still use magazines and other hard copy archival sources)?
• What pieces of data do you actually analyze? Just for grins and giggles, it’d take a team of about 200 people a full year just to read 1 Terabyte of Data… say nothing of analyzing it.
• Do you have a centralized data library that’s easily accessible to everyone who needs it?
• Does anyone “clean” (validate that it’s correct and in the right format) your data?
• How frequently does your team reference your data?
• How often is your data updated?
• How is your data shared? And is openly shared between different departments?
• Is the data enhanced or does it need to be enhanced with charts, graphs, etc., in order to be meaningful to the person reading it?
• Is your data accessible via mobile?
• What additional data would be useful/valuable to your company, its customers, or partners? Is there a way to acquire that data? Data can be acquired through sensors, devices, user interactions, and third-party sources/services.

The best news about Big Data is that it doesn’t involve any back-breaking work, at least you don’t have to move too much production equipment around. On the other hand, you’ll need advanced analytics to get any real value out of Big Data. Most of us are probably used to complex MS Excel sheets to compare one or a few data points.

1. Define your data needs. You should already have them, but you should also define your business goals on the basis of data points – the data that you are presently using and data that can be used to grow/improve your business.
2. Formulate a data strategy for your data/business objectives on how you will make use of your internal business-generated and performance data, social analytics, data mining/exploration, and/or third-party “Data as a Service” options.
3. Begin the activation of your data science team – probably starting with just one or two data/engineering savvy specialists to draft the technical requirements for your Big Data initiative. You can then scale up as needed, suffice that you don’t need to hire your entire data science department at once.
4. Unless you’re working at Area 51 or some other top secret project, you probably want to develop cross-functional, non-knowledge silo oriented teams. The easiest way to accomplish this is by creating a centralized data library wherein you can still manage permissions. If this sounds good to you, then you need to decide how much of your old data is still relevant (highly variable), how much effort and exactly how it can be ported into your new centralized data library.
5. Assess the difficulty, timeframes and expense of any software development (project requirements). that may be required.
6. Set up a long-term budget and plan to implement your data plan – start on training personnel, and establishing performance targets to evaluate project success.
   Continuous improvement can be added to both the automation and big data plans. These are pretty basic and will require customizing for your specific business.

There’s a lot more to digital transformation, automation, and big data. In our next post, we aim to cover more about big data “on demand” and how to distributed teams can leverage the success of your digital transformation.