In 2020, digital transformation is near the top of every CEO’s agenda. There are ambitious plans to use digital technologies, data science and AI. There is a modicum of dread as well — digital world seems to be different. In medium- to big-sized companies, say with more than 1,000 employees, implementation of any strategy takes much more than sound strategy itself. Any playbook must take into account the actions and motivations of the real people who would be working with it. That adds another layer of dread — you may start with a perfect strategy and still botch up the implementation. If that happens digital transformation can be very ugly.
In such situations, crafting certain values is always helpful. These are simple directives hiding profound intent. These simple statements are very useful in setting the direction for an organization as a whole, and help in making decisions in tactical situations where that direction is not very clear. In this article I have articulated some of the values that, in my experience, must underpin a digital transformation.
Many think about digital transformation as the process of taking customer workflows online. They are right. Having an online interfaces for ecommerce, ticket resolution, updating account information, etc., is the table stakes today. In truth, digital transformation is much much more, but you can’t have the iceberg unless you have the tip.
Digital transformation is done to an end-to-end process, not to a system. Many teams keep analyzing transcripts of conversations with their customers to identify the most prominent reasons customer are calling in, and automate the relevant workflow. The ability to dispute a credit card payment is my favorite. Most credit card companies let consumers do it online now. Earlier, this was a big reason for irate, long calls that led to negative NPS despite a positive resolution. The same principle applies to every aspect of your operations — where are people having to call others, and why?
There is a nuance I would add — you must think of not only your customer, but every user related to the company, including your employees. Calls between team members is good for camaraderie, but if it is a routine call as part of a repetitive process, it must be digitized. For example, in my work with pharmaceutical companies I realized that knowledge management is truly broken. Scientists have to call each other to find answers that may have been documented well, but are lost in the digital ether. A simple natural language search engine can save so much heartburn in these situations. The paper below talks about this issue more formally.
Digitally transformed many times does not mean completely digital. Many parts of many processes must go through human, or what the industry calls liveware. Thinking about such teams and digitalizing everything around them is especially important.
Then there are your vendors and partners. In the world of automotive design the concept of Product Life-cycle Management (PLM) has been around for three decades. It lets the manufacturer, say General Motors, share its designs to its Tier 1 suppliers, say Magna, which, in turn, can cascade them with its suppliers, say Sundaram Fasteners. So any design change that Sundaram makes is visible to GM in real time. They can approve it, comment on it or change their own designs around it. Digital transformation is a great opportunity for every industry to adopt this concept of tighter integration with their vendors and partners.
Digital transformation is not about technology, it’s about users. This is why a product manager must lead such efforts. The north-star for any digital transformation initiative must be to minimize the effort a user makes to do something with your company. If a user can just tap on their phone to get something done, instead of calling, that is a huge victory by itself. Still, that’s only the beginning. A successful digital transformation program asks two more questions.
First, can we redesign the process to make sure that it is the simplest possible? Let me give a counter-example to this. Try creating a ticket for some issue on LinkedIn, even if you are a premium member. The link exists, I promise, and once you create the ticket the LinkedIn customer service team does respond within 24 hours, but it is very difficult to find that link. My friends at LinkedIn tell me that it may be a deliberate choice meant to optimize other things. However, compare this to the multitudes of platforms that have a chat icon right on the bottom right corner of every page. Even there, I prefer systems that respond immediately, even with an incorrect answer (and an option to escalate), to those that make me wait for a human.
Digital transformation is not about technology, it’s about users. It’s done to end-to-end processes not individual systems.
Second, can we design interfaces, mobile-first, of course, that unclutter every choice the user has to make and transform even a complex process into a series of small choices? My favorite here is how Turbotax takes the process as complicated as taxes down to series of intuitive yes/ no or multiple choice questions and fields to enter very specific values.
1. Conversational AI: Code/No Code
2. Chatbots 2.0: Simplifying Customer Service with RPA and AI
If there is no easy way to do this for everyone, I advise my clients to minimize for maximum number of people. A frequent technique I suggest is to have only two or three choices available to a question, and hide the other 17 under “More.” Google takes this to an extreme — the search engine is actually very powerful in faceted and range-bound search. The advanced search options within Gmail and enterprise versions demonstrate that. However, a vast majority of Google’s users don’t need these features. So, it has not bothered putting them on its home page. There you can do only one thing, and that is not advanced search.
One last thing — don’t have standards for your customers that are different than those for your other users like team members, vendors or partners. We all are humans. Plus, it is easier to maintain one set of consistent features than multiple, as most operations people would agree.
In the digital world, humans should review, not do. This does not take control out of human hands, it just make their life easier. Let’s consider the multiple levels of this paradigm.
If there are decisions that machines can guess, a digital process should make those decisions and present to humans for approval. For example, Facebook’s AI automatically identifies your friends in a picture you upload, and then prompts you to tag them. The right balance between autonomy and human control is very important, and it’s your job to find it. As per an article in New York Times, linked below, the Chinese government takes the same technology and goes ahead to tag everyone at will. Most think this is a step too far politically. However, the user for this digital transformation is not the populace. It’s the police.
Many times the machines can assist in narrowing down the options for the user. I have written earlier in the article about hiding unnecessary details under a “More” tag. Simple analytics can help decide which option to feature and which to hide. As a result, most people find the interface very easy, and yet the users who need the uncommon options can click once to access them.
Even if the decision has to be entirely human, digital transformation calls for collecting and presenting all the necessary evidence so that humans can make those decisions easily. A good example of this is AngelList, the online platform for startups. If you post a job there the platform asks you for various details like salary, equity, etc. Next to each question there is a range that other people have used for similar roles in similar geographies. The user has all the evidence to make their decision.
Such thought does not always need sophisticated analyses. True, there are advanced algorithms, like those based on neural networks, that show a lot of promise. However, thoughtful application of this value based on available data can get great results with simple heuristics as well. The important thing is to focus on the user and minimize their effort. Below I have linked another article that delves further into issues related to choice of technology. It is focused on one aspect of digitalization — chatbots.
Regulatory compliance touches almost every industry in some form. Finance and Healthcare are regulated per se. Laws like GDPR and CCPA put restrictions on almost every other industry as well. Even if there is no legal requirement, a comprehensive digital transformation program must build on certain internal policies and standards about data, interfaces and security.
This value simply says that the digital transformation team must integrate compliance with regulations and policies right from the outset. The transformed process must monitor data flows stringently, clearly define the right access/ security policy and make copious records, likely in the form of logs.
In my experience, compliant processes are easier to manage, get budget and regulatory approvals quicker, and can better act as a foundation for other processes. Over time each process adds to a faster pace of digital transformation. A non-compliant process only adds to confusion.
If the focus on users is the most important success factor for a digital transformation project, management of data involved in the transformed process is a very close second.
Let’s look at the life-cycle of this data. It is often created from sensors and automated devices. A significant amount is volunteered by customers, vendors or partners. The data is stored in some repository. In well-managed organizations it is stored only once, others copy it multiple times (and lose track). The data changes forms multiple times and gives birth to a lot of derivative data. Then it never dies. Over time it becomes less and less relevant till it’s shipped off to some permanent storage, lost forever.
Digital transformation does not change this life-cycle directly. However, to ensure that any data, once generated or captured, maintains its integrity throughout its life and is seamlessly available to all legitimate users of that data, is a foundational value for digital transformation. Yes there are areas that you can digitize without a comprehensive data framework, but very soon you will have to come back and do the work again unless you have thought about data life-cycle end-to-end.
Here are some best practices with regards to data-management and digital transformation that I have learnt working with the best teams:
- Never design a digital process that relies on typing in data that is present elsewhere. Instead, connect the sources of data to your process using APIs and ETLs. This is also sometimes referred to as Straight Through Processing.
- As an extension to that never ask for the same data twice. If you have the data, pre-populate the relevant fields for confirmation. It is important to not have multiple versions of the same data.
- Whenever a process captures or generates data, put in real time mechanisms to ensure its quality. Not-In-Good-Order data is a problem big enough to merit its own term — NIGO. Unless corrected at the point of capture, such data will render the entire downstream process useless.
- For every bit of data captured or generated by a process, maintain a master repository or the Single Source of Truth. There are many techniques to get to a good level of integrity without heavy investments in technology.
Some readers may be thinking at this point that their organization is not sophisticated enough to embark on any ambitious digital transformation journey. The good news is that data management takes more thoughtfulness than dollars. You should still chase your ambitions. Just that make data management an inherent part of that journey.
There are tomes written about agile, not only as a software development methodology, but as a mindset for product management, process design and project execution. I will not try to compete with that literature in a paragraph, except to reiterate and ask you to measure everything, review everything, and change ruthlessly.
These values have emerged from my ~10 years experience helping mostly Fortune 500 organizations with digital transformations. Still, we have to agile about these values as well. Do let me know of your thoughts.
