Making the most of an industrial PhD

There are many advantages to doing an industrial PhD. You build important relationships within industry; have experience of working on real problems which can improve your employability; often have lots of data to use; and may even be offered a job on completion of your study. There can also be a number of challenges. This blog post identifies common challenges and calls for an industrial PhD framework to be developed.

When I first got my job as a lecturer in Business Analytics, a fellow PhD student got in touch and asked me how I had managed to get such a prestigious position on the back of an industrial PhD, without (at the time) any top tier academic publications – an often critical aspect of academic job applications. He was saddened by the fact that he was finding it difficult to publish in a top tier journal with his research and he felt the reason behind this was because he was working on an industrial PhD. He told me that his supervisor had very few top tier articles because she was always working on case studies in industry and he worried that this may limit his options if he wished to work at a University, rather than continue in industry. I understand the difficulties of trying to achieve top tier publication with industry-led research, having faced some myself, and I have heard this complaint from others.

Part of the problem is the requirement to publish in journals which have a high bar when it comes to applied research. One has to perform multiple or longitudinal (over a long period of time) case studies to be able to prove that your finding(s) could be applied to more than one industrial case. On an industrial PhD you often work for a single company, over a fixed period of time, meaning your research is inevitably tailored to their needs. Your findings could potentially be applied to another company, but evidencing this would most likely fall outside of the boundaries of your PhD and take more time (a limited resource in doctoral research).  However, if we can’t generalise our findings how can we say that we have found something useful outside of this one industry?

Then there is what the industry wants - how often will an industry be a good test bed for your ideas or be happy for you to ‘experiment’ on them?

Industrial studies can easily become something which looks more like a consultation exercise than an academic one. In practice (industry or otherwise) business-focussed PhD’s tend to come in two flavours:

1. We have a problem or are making developments and we think your research will help us to solve/do it
2. We have many problems and have no idea what to do about it.

Both of these can come with difficulties and opportunities. In the first instance, a company will know what they need and if they have approached an academic institution are likely to be very engaged. If an academic has approached them, then the industrial partner may be sceptical and may never have worked with academics before meaning they most likely are unfamiliar with what to expect from such a partnership. My PhD was of the second flavour. In this area, there is scope for ‘blue sky research’ but there is also a pit of despair!

I was lucky in my PhD to be studying in a company that was completely engaged in the research objectives and which desired the opportunity to develop new and novel techniques for re-engineering their processes. However the difficulty of pleasing two masters was ever present during my studies, which led me to examine this aspect of industrial partnerships with academia. In my opinion, the main issues which can cause difficulty when studying in industry are:

• Ethics
• Disclosure
• Different pace, goals and value models

Ethics

In general, academia is more concerned and more rigid about the ethical aspects of research, but when you are working within an industry you are often bound by their own ethical code. In my case the industry was very highly regulated and the ethical emphasis was on maintaining this.

Much of my work used qualitative techniques. Qualitative techniques are about listening to what people say and watching how they behave, such as problem structuring methods and action research. Problem structuring helps a researcher to help organisations to explore their problems in a collaborative way, it has many similarities to well-practiced lean exercises such as Continuous Improvement. Action research also explores problems then puts in place an agreed intervention which the researcher can observe and study. They then recall their observations with the stakeholders in the organisation and a next process of agreed intervention takes place building on the last one. These methods allow a free and open exploration of problems without assuming that people will change their minds or can be told what to do by research or by the boss!  Ethically these can be useful techniques as the methods make the consideration of free will in participation, and agreement upon interventions, explicit.

Disclosure

I, and the University where I was a student, had to sign a Non-Disclosure Agreement (NDA) meaning we are liable if anything is said, in relation to the research, which the company deems damaging (The University of Leeds has approved templates that are used in such situations and advice should be sought from your Graduate School / Faculty Research Office). Any papers I want to publish, or data I wish to use for teaching, must be agreed to by the company. Even my PhD thesis can only be seen by researchers approved by the company first. I also had constraints on any recorded information and all of the data I gathered needed to be kept securely in line with the terms of the NDA.

This may seem harsh to anyone who has not had dealings with industry but many people who have worked in the commercial environment will understand these necessary constraints. In my case, I was able to build trust by showing the company executives all of my writing and presentations, which they soon realised did not disclose anything harmful. I have just had a paper published in an international journal and my industrial supervisor found reading and approving this paper a worthwhile experience. He was happy that all commercially sensitive data had been disguised should anyone make the connection, even though I used a pseudonym for the company.

From an academic perspective, NDAs can be constraining but don’t have to be completely restricting, and the terms of such agreements require that you are careful and thoughtful with data. Despite the restrictions, the first three publications that have come out of my thesis are in the academic domain, either as working papers or full publications.

Different pace, goals and value models

Possibly the most disruptive aspect of an in industrial PhD is the different pace of development and delivery which is driven by different value models and goals. Perhaps the easiest way to explore this is through a comparison table (keeping in mind that these are generalised working structures).