Three years ago, during a training on Process Mining by prof. Wil van der Aalst, I learned that “data was to become the new oil”. These days, I observe the increasing confirmation of this statement.
The IoT (Internet of Things) produces huge amounts of useful data and this will – if not exponentially – further increase in coming years and decades. Nowadays, nearly as much data are produced on 1 day as all data produced since the beginning of time until 2003. By 2020, it is expected that 50 zettabytes of data will be available.
Add to these huge amounts of data the increasingly smarter technologies to analyse and to exploit them – like Artificial Intelligence, Machine Learning, Deep Learning, etc. – and you can imagine the potential all this offers to improve businesses, and obviously also business processes.
It should not surprise you anymore to see the increasing existence of data brokers.
Hence, each organisation – either for profit or not for profit – should seriously consider (big) data and (data) analytics as strategic assets.
3 main ways to get value from data
Before diving into the Operational subject, let us see how an organisation can look at big data to generate even more value.
1. Making better decisions
Data availability and data analysis may be the source of information leading to improved market and customer intelligence. This can be for example through better insights like
- What customers expect and want
- Which products & services customers use
- How they consume these products & services
- How customers evaluate those products & services
Such inputs and feedback enable businesses & organisations to improve their sales & marketing efforts. This article illustrates how Heineken applies the so-called ‘social CRM’ and how it uses social media as a basis for future customers.
It is even not (only) about making better decisions for the organisation itself, but also for its customers. Think of bike or scooter sharing : sharing vehicles used to be at well-fixed points in a city. Thanks to geolocalisation, customers can now get a dockless (e-)bike, (e-)step or (e-)scooter at their fingertips, i.e. any place where the previous user has left one. Through your smartphone, as a client, you can then look up where to find the nearest vehicle to use for your trip to elsewhere in the city. The client’s decision thus depends on the availability of bikes or scooters he gets as data through his smartphone.
The US-based restaurant chain Dickey’s Barbecue Pit goes even further. It has developed an own data system to get even better insights so to increase its sales and to improve operations, training and even menu development. Their system collects data in near real-time from all the ‘point-of-sale’ (say check desk) systems, from marketing promotions, though also from loyalty programmes, customer surveys and even from their inventory systems.
Combining all these data together allows them to respond “on the fly” to beneficially influence supply and demand. This way, it is possible to send a text message to customers of a local area to boost a lunchtime when they see that there is a lack of customers at a certain day.
2. Improving business efficiency & operations
Data enable efficiency improvement in any (functional) department of organisations. Even not only regarding internal processes “within its 4 walls”, though also for inter-organisational processes like logistics or Supply Chain Management.
Also supporting processes like HR activities consume data from social media. Think for instance of LinkedIn for recruitment purposes. Not to speak about facility management, where maintenance costs may largely be improved thanks to sensors enabling predictive maintenance. Which helps diminishing rather expensive preventive maintenance, while avoiding costly breakdowns – often the cause of too long downtimes and high repair costs.
3. Monetising data & information as such
This is about companies that collect and analyse data to deliver insights to their clients. Insights such as customer satisfaction through analysis of social media, or forecasting patterns by analysis of searched words or expressions like Google Trends.
Or any data – or information – broker on very specific topics. Take Enhesa as an example, a service company that scrutinously screens any new regulation on Environment, Health and Safety around the world, to help its clients – often multinational companies – to be or to remain compliant in these 3 domains.
Another example is John Deere, which sells data on machinery performance, though also on soil conditions, and other kinds of data enabling farmers – its target market – to improve the yield of their crops.
Optimising your business processes through data
Let’s now zoom on the use of (big) data for operational improvement, say data driven process improvement.
1. Improving individual tasks (the process building blocks)
Business processes exist of activities or tasks, say the process steps. That’s why all of these process steps should thoroughly be analysed on how data may optimise them individually.
Read in this article how the Tata Steel factory of IJmuiden (the Netherlands) succeeded in considerable improvement of raw iron heating. Thanks to data analytics, they eliminated costly waste – caused by lost liquid raw iron boiling over – during heating. A nice example of better operational control isn’t it?
Moreover, they also improved the heating activity by heating the iron much faster. Two birds with one stone ; in this case, an algorithm which analyses sixty different parameters each second, so to calculate the probability of the liquid iron boiling over within the next minute.
Such algorithms and the use of big data do obviously not come “out of the blue”. It is most often the result of looking at how data may help to
- increase the speed of a task or other said decreasing the time needed to execute it.
- decrease any other type of (lean) waste, including any resource
- improve quality for the “customer” ; in the case of a task, the customer being the next process step.
- improve the overall business process, its output(s) and objective(s).
- make it easier and even more motivating for human operators to carry out the task. E.g. making it even more ergonomic, healthy, intellectually challenging, less risky (safer & more secure), etc.
2. Optimising entire processes
All above mentioned principles for individual Activities or Tasks are obviously also applicable for the entirety of all these tasks, thus for the process(es) which they are part of. Though as a business process is more than the sum of its steps, you should also consider following additional ways of optimising them.
Interactions between tasks within a process
Even when the tasks are optimal, the process itself may be sub-optimal or inefficient.
Look at the “spaces” between the tasks of your processes : are there any wastes (including wastes of time) with the transfer of goods, labor, …? Particularly when consecutive tasks are carried out by different persons or teams. Maybe these wastes are even not known yet, and the use of specific sensors might make those wastes visible.
Tip: if you use process models including swim lanes, it is always interesting to analyse how materials, data, information or knowledge are transferred from one person – or one team – to another. These are often sources of data driven process improvement, indeed.
Example: Imagine an administration that issues building permit files, while the final decision – and file – depends on inputs from several other persons or teams. Isn’t it convenient and time-saving when the person responsible for this final step is aware of the status of all inputs s/he needs? For instance by receiving a signal as soon as all these inputs are available for the final task.
The same applies, of course, for product assembly tasks when this final assembly depends on the availability of all pieces to be assembled.
Raise the bar for your process objectives & process output criteria
Look at how you could exceed, increase or even redefine process objectives and process outputs thanks to the availability of specific data.
If the current lead time of your process(es) to manufacture a product, or to deliver a service, is longer than what customers expect, why not analysing where in your process(es) data could help to eliminate time wastes?
Or if your process is still subject to too many non-conformities leading to rejected products, why wouldn’t you consider data or sensors to find the (root-)cause(s) of these non-conformities? If you doubt how realistic this is, then please read the paragraph “automatic (root-)cause analysis through intelligent evaluation” further in this blog.
Integrate IoT data within your business processes
This means using IoT data directly in your business processes, so to improve the effectiveness and/or the efficiency of your processes.
Assume that your organisation is supplying time-based services, and that only the time for which your employees are at your customer’s premises is billable. In the past, employees should have written down – or in the best case, they could have registered it on a tablet -, each time they arrived and each time they left a customer. This was not only a tedious task for your workforce, but it was also subject to many risks :
- risk of errors – particularly when people forgot to note times and had to remind those times at the end of the day,
- risk of losing the paper with all the (time) data,
- risk of fraud – e.g. by claiming they passed time with a customer while doing anything else,
- risk of many discussions with customers on the time precision
- etc.
This now belongs to the past, as Telematics enables you to register all these times automatically, even in real-time. Though this goes even further : importing all these (time) data directly within your invoicing software allows you to invoice your customers in a much more automated way and moreover based on very accurate times. This Telematics service is actually one of the solutions which GeoDynamics offers to its clients.
Analyse your past process instances and event logs with Process Mining
If you do not know (yet) what Process Mining is about, then please have a look at this blog – and the 2 following ones which describe how to get even more value through applying Process Mining.
Because Process Mining depends on available data, namely event logs, you obviously need those data before you can improve your business process through Mining. A frequent reaction when presenting Process Mining is “nice technology, but where should I get those event logs from?”
The rather simple answer is : “if you do not have them, get them”. Making these event logs available is no rocket science these days. Indeed, in this world of IoT (Internet of Things), with all kinds of possible sensors, you do not depend (only) on expensive and rather rigid information systems like an ERP or CRM or whatsoever to get thorough insights on how your processes are really run. Think of any type of sensor that may help you to fill gaps in your process reconstruction.
Imagine that you are an actuary, who is looking to continuously improve insurance premium calculations, according to driver’s risks for accidents. Referring to the example from previous blog, i.e. the car insurance company using Telematics to improve the management of risks. It is more than plausible that these data could also be used to continuously improve actuarial processes, isn’t it?
Automatic (root-)cause analysis through intelligent evaluation
Intelligent systems which do not only evaluate the quality of products, though also automatically analyse and identify the causes of poor quality in business processes are becoming increasingly popular and are used more and more today.
As this article illustrates, data analysis techniques allow to identify and even help to localise error causes in a manufacturing process. This way, an early warning system can be foreseen, so pointing out potential sources of errors. And this obviously helps to minimise non-conformance costs, possibly to zero.
3. Optimising the overall supply chain
It goes without saying that the same principle for process improvement can be applied across the entire supply chain, as long as supply chain partners trust each other enough to exchange data. According to IBM, 65% of the value of a company’s products or services is derived from its suppliers and its supply chain.
You may remind the bullwhip effect explained in this blog. Well, sharing data from the 1st to the last actor in the supply chain could help to eradicate – or at least to considerably decrease – this bullwhip effect and to very significantly increase the overall supply chain efficiency.
Asset (e.g. goods) tracking and connected fleets enable you to have a view over the entire chain at any time. Partners who collaboratively want to improve the overall supply chain – so to deserve the end customers even better – could also apply process mining principles on the entire supply chain, as they would do for their own processes.
This inspiring pdf – even though a few years old already – illustrates the actually illimited potential of the IoT and data analytics in Logistics.
4. Some recommendations
Through a survey of 300 IoT-practicing businesses, McKinsey identified 9 best-practices which separate IoT leaders from IoT laggards.
- Do neither think – nor act – (too) small : as the first use cases represent a learning curve – and thus an initial investment with rather low return -, the highest return will occur once you will have acquired more experience with applying IoT and analytics. Hence, even if you start with pilots, consider it on a large scale from the beginning.
- BPM as a lever for IoT’s value : even though it is about technology, the IoT value will (mainly) be generated by process redesign and the resulting improvements.
- Consider and use advanced technology and end points : do not consider IoT as such only ; look at other recent techologies – e.g. the several forms of artificial intelligence, augmented reality, virtual reality, etc. – and possible combinations of those with big data.
- Determine & define how IoT will create value : look for strong business cases and how the IoT could solve your customers problems even better than any other value proposition.
- Involvement and commitment of the C-level, particularly the CEO : like any business transformation, it needs the c-level support, especially from the CEO. As it is about process (re)design and because many processes are cross-functional, the CEO will be the best referee to make sure that the investment provides the most optimal return for the entire organisation.
- Contribution from the whole organisation : like for any effective strategy implementation, you need the support from all departments and all levels within the organisation.
- Start from existing products & services : though the IoT may enable to create new business models, new products or new services, you yet best start from existing offerings.
- Dare to use third-party IoT platforms : don’t be overambitious by doing everything yourself, though rather look for ecosystems and/or partners who already acquired top experience and knowledge of the IoT and big data. This will help to fulfill the above mentioned recommendation n°1.
- Be prepared for cyberattacks : unfortunately, the IoT also have risks. Apply a sound risk management to mitigate those risks, without falling into doom thinking.
Please share your experience or your thoughts about the use of data – or data analytics – through below Comments box. And receive in return an e-book from a very business minded big data expert, illustrating the power of using data in many industries.
P.S.: If you found this information useful, don’t hesitate to share it with your Facebook friends and fans, LinkedIn contacts, Twitter followers and Google+ circles, through the share buttons below. Thank you!