IoT in pharma manufacturing changes company culture

Digital transformation comes with unforeseen yet sometimes very beneficial consequences. Who would have guessed that introducing IoT (Internet of Things) to pharmaceutical manufacturing could have a broader transformational impact on a traditionally conservative company culture?

Conservative Pharma Industry

As a bit of background information, there are many reasons why the pharmaceutical industry tends to be more risk-averse than others. Here are some key considerations:

  • Long-term investment:
    Developing an innovative new drug easily takes 10 years and costs $2.6Bn upfront (according to the Pharmaceutical Research and Manufacturers of America, PhRMA) before the actual product reaches the market. Pharmaceutical development remains a high-cost, high-risk business where mistakes are punished harshly and can ruin a company.
  • Regulated industry
    Regulatory authorities, such as the Food and Drug Administration (FDA) in the United States and corresponding agencies around the world, closely inspect every aspect of the development, manufacturing, and marketing of medicinal drug products from pharma companies. If a company is found out of compliance with Good Manufacturing Practices (GMP), severe financial penalties can be imposed and drastic consequences loom – including shutting down the business altogether.
  • Human lives are at stake
    Pharmaceutical manufacturing is where the rubber hits the road: Any problems in the manufacturing process can easily affect product quality and thereby directly threaten the health and lives of patients. Every facet must be closely observed, and regulatory inspections are frequent and thorough. Therefore, changes to the manufacturing environment are done most reluctantly by companies to minimize risks.

The burden from these limitations weighs heavy on the organization and lends itself to a conservative mindset and cautious approach. Change is not always welcome as it induces risk that could jeopardize operations and outcomes.

Moving to IoT

The Internet of Things (IoT) is more than just a bunch of devises and sensors that communicate with each other and generate a constant stream of data: IoT affects not only how we (make things) work but can also affect how we think and the foundation for our decision-making.

The traditional process in pharmaceutical manufacturing produces batches of product. It requires many human process steps from preparing and calibrating machinery, running the batch, examining the quality and then cleaning and preparing the equipment again for the next batch of the same or an entirely different product. During the process, devices collect data in their own ‑often proprietary‑ data formats that may be hard to access. The data has to be collected, combined and interpreted in a time-consuming process full of interpretation barriers and prone to human error. Even worse, “over 70% of the data in manufacturing is never touched” according to the CEO of Bigfinite, an IoT provider, and certainly not timely. This comes at a cost as this example shows: An American pharma company reportedly lost $20 million worth of product when a $3,500 vacuum pump broke down.

Around 30% of the Top 20 pharma companies started introducing IoT in their pharmaceutical manufacturing (according to GEP, a supply-chain advisory firm) to enable faster and continuous data collection from several processes for real-time monitoring, integrated analytics, and more timely decision-making. The paramount goal was to meet regulatory demand, such as the FDA requirement for continued process verification.

What comes with IoT

However, IoT relies on Cloud computing to provide digital connectivity across the entire supply chain from production to market and across plants. IoT Cloud computing may come with the necessity to use third-party-run servers for data storage and calculations raising the all too familiar fears of pharma managers and employees. Often enough it is the employees who interpret regulatory guidance to narrowly and don’t dare to rock the boat by changing the current GMP (cGMP) out of inflated data security concerns and the doomy risk of falling out of compliance.

While care certainly needs to be taken when implementing the new technology and while processes need to remain compliant, the FDA has already shown flexibility and set a precedence in approving the shift from batch to continuous manufacturing for Johnson&Johnson’s HIV drug PREZISTA.

More recently, the regulatory concern no longer seems paramount. Instead, management understands that IoT opens the door to massive and much-needed cost savings, shorter cycle time, right-sizing operations, increased productivity and higher competitiveness in the highly competitive pharmaceutical market arena.

People transformation beyond digital

Interestingly, all these more technical aspects can distract from how IoT in pharmaceutical manufacturing can lead to a broader shift of mindset throughout the organization:

Sharing and compiling formerly compartmentalized data across different parts of an organizational practically breaks the well-established and well-protected silos in many organizations. Suddenly, everyone seems connected to everyone else in the company and departmental borders fall while the process becomes visible and more transparent in real-time.

The fundamental shift with IoT and Cloud computing forces management and workers to adapt to the new technology and to connect with others outside their immediate organizational silo. The newly integrated informatics can include Enterprise Resource Planning (ERP) and financial systems. Sharing the data trove happens not only within a manufacturing plant but also across 25 plants at Pfizer, for example.

The technology-induced visibility and management of the manufacturing process challenges the traditional mode of operation and encourages employees trying out something new. If managed well, this mindset shift can be used to crack the barriers and drive a favorable cultural change throughout the organization. It enables but also pushes employees to continuously improve manufacturing operations while it also translates and proliferates into all other aspects of their work.

Summary

IoT technology in pharmaceutical manufacturing not only improves the productivity and competitiveness while maintaining regulatory compliance but also challenges and steers employee mindset away from overly conservative restraint toward collaboration and continuous improvement – and thereby shifts the organizational in favorable directions.

Advertisements

Intrapreneurship Case Study at the Foreign Trade University (FTU), Vietnam

Students of the Creativity, Innovation and New Value course at Colorado State University discuss the popular teaching case study “Boehringer Ingelheim: Leading Innovation” (available at Harvard Business Review (HBR) and the Ivey Business School) with intrapreneur Stephan Klaschka.

Intrapreneurship Case Study at the Foreign Trade University (FTU), 

As a cooperation between Colorado State University (CSU) and the Foreign Trade University (FTU) in Hanoi, Vietnam, FTU students are tasked to develop an opportunity in a team and conduct a feasibility analysis on the opportunity that they present to the class on June 4, 2018.

I will join the students and visiting CSU professor Robert Mitchell for a live discussion of the teaching case study Boehringer Ingelheim: Leading Innovation, which is available at Harvard Business Review (HBR) and the Ivey Business School,

In this teaching case study, the case writers Professor J. Robert Mitchell, Ph.D., and Ramasastry Chandrasekhar, of Ivey Business School, follow the footsteps of Stephan Klaschka’s intrapreneurial approach to innovation within a global pharmaceutical company (FORTUNE Global 500, Top 20 Pharma).

This intrapreneurship teaching case study is used by staff and students of Intrapreneurship and Innovation in business schools around the world. and features my career as an Intrapreneur at a major pharmaceutical company.

HBR-logoIvey-logo

Measure your company culture in real-time!

It is difficult if not impossible to assess organizational culture directly.  Instead, managers favor surveys to measuring organizational climate as a first step.  However, surveys fall short in many ways and can lead to skewed results as input to managerial decision-making.  Better than surveys is observing employee behavior with a meaningful metrics.

What is your organizational culture?

No matter where you work, you are a part of it:  the organizational culture.  Culture is understood to comprise shared beliefs, values, norms, traditions but also myths of employees about interpersonal relationships, behaviors and activities of the organization.

A (favorable) strong culture indicates alignment to organizational values and goals – some call it the organization’s personality.  This is the internal glue for collaboration and outstanding results as an organization.  In a strong culture, ‘can do’ stories share ‘how things are being done around here’ that inspire and motivate employees to action and ‘organizational citizenship behavior’.  A strong culture supports employee satisfaction and retention as well as innovation and productivity. (See also: How to create innovation culture with diversity!)

In contrast, misalignment of values and goals in an unfavorable weak culture has an eroding effect.  They easily lead to extensive rules and bureaucracy that rely on exercising control.  Working in this place is not much fun.  Don’t expect anyone to go the extra mile!

Unfortunately, organizational culture is a slippery and complex subject, which makes it hard to grasp – and hard to measure directly.  It is easier to feel than to express.
– Try it!  How does the culture of your organization feel in your gut?  How about putting it in words?

How to measure culture?

A common approach is to measure a company’s organizational climate by looking at the culture’s outcomes or consequences rather than trying to grasp culture directly.  Thereby, the climate is used as surrogate marker for the underlying culture, since outcomes are easier to observe and to measure.

Here we find a handle on whether the employees are happy at work and feel valued, if they enjoy their work environment and trust their colleagues, if they go the ‘extra mile’ for their team – or if they are frustrated, disengaged or even act hostile against coworkers or the organization.
Factors to establish a metrics offer themselves relating to –for example- communication, accountability, behavioral standards, rewards, trust, and commitment.

Organizational climate’s primary driver is daily leadership that influences the expectations as well as the behavior of all individuals in the organization.  The leadership also determines the organizational structure, another key to an organization’s effectiveness.  Both enable the organization to reach its goals, but also reflect priorities and heavily affect how employees communicate, collaborate and interact with each other.

Many factors obscure the clear picture including rapidly growing workforce and geographic separation but also the way we actually measure organizational culture.

Yet another survey?

Many companies invest in surfacing climate data to ‘feel the pulse’ of their staff and to confirm positive effects or apply corrective action to adverse findings.

The most common way to measure climate is a climate survey and repeated to compare changes over time.  Despite our daily information overload, many companies typically use surveys to collect data from as many employees as possible to paint a representative picture of the company.

Surveys seem the first tool in the managerial arsenal.  They appear attractive, seem simple and powerful.  Survey results are seen as straightforward, clear, quantifiable and reflecting the ‘truth’ since the workforce was asked directly.

‑ But are surveys truly the best tool available or even an proper tool at all as a starting point?

What is wrong with surveys?

Unfortunately, surveys are far from ideal for several reasons.

The first issue we face is that there is no common standard for measuring the ‘climate’.  Every organization or consultant comes up with a different scale.  If an organization introduces its own scale and applies their metrics consistently, it can build a database over time.  The data, however, only compares directly against other client organizations or industries that were measured similarly, i.e. sharing the same scale, at a premium for this proprietary benchmarking.

Even worse, results hardly compare because surveys ask questions relying on language.  A slight nuance in phrasing of a question may change the meaning and influence the responses.  After all, words are ambiguous and open for interpretation – and even more so in a multi-cultural society and multi-lingual.  For consistency and easy processing, they typically come with a fixed set of response options such as multiple choice, which can limit the responders’ options and influence what they respond.

Often overlooked, the real workload comes after the survey closed in the analysis, when you start slicing the data to combine questions, sub-populations or start exploratory analyses in an afterthought with all the shiny data you find in your hands that seem to open endless opportunity for finding answers.  This is where you easily run out of time or budget – and where it becomes tempting to cut corners just to finish up and deliver results while sacrificing depth and consistency.

Surveys tend to be inherently skewed – Why?

When was the last time you enjoyed taking a survey?

Our email in-boxes are full of customer service surveys for a recent purchase or some service call over the phone or online.  The whole world seems wanting to improve their services – and sends us a survey.

However, surveys are far from ideal for several reasons including these (and many more):

  • Fatigue – There is no shortage of surveys these days.  Coming back to our information overload and time constraints, many people just don’t want to fill out another questionnaire or find time for it in the first place.  ‑ Did you ever give random responses or skipped questions just to get it over?
  • Privacy – Some other questions you may not feel comfortable answering in the first place because they invade your privacy by collecting data with questionable benefit to you.
  • Anonymity – in the computer age, anonymity is hard to find.  Even in an otherwise anonymous survey, the combination of responses can identify individuals under certain circumstances feeding privacy concerns.
  • Past – Surveys measure the past.  Even the most credible survey questions inquire about past behavior at best, which is the most solid data you can get out of a survey.  The results may be good for forensics but hardly reflect the current situation.
  • Diversity – a diverse workforce can come with communication barriers of language or cultural background that leads to misunderstanding. Geographic idiosyncrasies can induce further bias in distributed organizations.
  • Delay – surveys take time to prepare, to conduct and to analyze.  Don’t expect to get the results anytime soon, especially because you cannot control when your responders choose to respond.  You have to adjust to their schedule, so getting survey results removes you far from ‘real-time’.
  • Precision – in surveys, you can easily measure everything to a dot and even farther right of the decimal point.  Some give you the tendency to ask and measure too much just because we can or we feel the results (and our work) look more credible this way.  Often it is an illusion that a higher level of precision adds to clarity when it adds to inertia instead by a flood of obscure information irrelevant to the decision you want to make.

The list goes on… you got the point.  The question remains what is a better approach to measure organizational climate?

Why it is better to measure behavior

A survey measures our intent – not our behavior.  Unarguably, behavior is a much stronger indicator than intent.  It comes down to whether we observe people putting their money where their mouth is or if we get only the lip service that a survey represents.  – Think of it as the litmus test you remember from chemistry class: It shows you the truth and reveals whether your assumptions hold true!

Let us look at the benefits of measuring behavior using the same list again:

  • Fatigue – As human beings we can refuse to respond to a survey ‑ but we cannot stop behavior as such.  Even if we refuse to respond, this is our observable behavior and becomes measurable.  For example, if large parts of the surveyed staff do not respond to the survey, this tells you something about the organizational and what is important to the staff.
  • Privacy and Anonymity – Usually, your observable behavior as an employee is not a privacy concern, since you are out in the open and visible to your co-workers anyway.  Again, you cannot not show behavior once you agreed to go to work, there is nowhere to hide. 
    (Let’s not derail by focusing on or encouraging questionable, unethical or even illegal intrusion of privacy at the workplace or outside.)
  • Past – Our observable behavior is now, it is the present.  You can’t get better real-time data!
  • Diversity – For observations, it does not matter if your workforce is diverse or understands the questions you ask.  There are no communication barriers when it comes to observing behavior. Actually, quite the opposite holds true: the employee behavior can help you to better identify communication barriers or other issues that a survey would not reveal!
  • Delay – observing behavior also takes time but it is mostly the time to identify what you want to observe for what reason as well as observing it and then summarizing the results.  There is no polishing questions and response options.  You get to results faster because you are on your schedule and do not have to wait for responses trickling in.
  • Precision – key is to measure only as much as needed, i.e. to establishing necessary and actionable facts.  Forget the fluff and focus on the one or two most important aspects needed for effective decision-making.

How to measure behavior?

Now, measuring behavior is not always easy.  It requires thinking through the cause-and-effect dependencies.  – A well-known example of how not to do it is the questionable relation of using the price of butter in Bangladesh to predict the stock market in the USA…

What the right metrics is depends on what you want to find out.  What is the underlying business problem you are trying to solve?  Many roads can lead to Rome, so to speak, but the basic idea is to keep your target simple.  Choose a target that is meaningful, robust and easy to observe.

Clarity helps.  As much as we crave being informed and gather data this approach is not helpful, since it tends to produce clutter.  Instead, focus on measuring the minimum you need as the basis for making a sound decision.  Don’t fall for the nice-to-have and garnish data you could have in addition.

How precise do you need the results really to be?  – As an example, you may be concerned about low meeting attendance.  Does it make a difference for your decision-making if you find out that in three consecutive meetings “63.26%, 58.18% and 69.4% of the invitees did not show up” versus “on average, 2/3 don’t attend”? – Let me guess, “2/3” does just fine to decide slimming down who is invited in the future or to change the purpose of the meeting, right?

The key is to stick to clearly observable behavior.  Some solid behavioral data may already exist within the organization.  – For example, a long tenure and low turnover may reflect that employees prefer to stay with organization, while many internal job applications reflect dissatisfaction with their current position or department.

Bottom line

Next time you think of running a survey consider taking a close look at employee behavior first!

References