Machine learning: Unlocking the hidden insights inside | NTT DATA

Wed, 29 January 2020

Machine learning: Unlocking the hidden insights inside programme status reports

Large-scale transformation programmes are almost inevitably riddled with unwanted surprises and challenges; in fact, the failure of IT projects in the USA costs the economy up to $150 billion a year.

Understandably, methods of effectively tracking programme delivery are in high demand, and numerous consultancies have developed their own advanced analytical techniques to make predictions by analysing reporting trends.

These tools assess factors such as RAG status, date changes and risk ratings, analysed against a huge backlog of project data to offer predictions and insight. However, status reports provide much richer data than just quantifiable metrics.

The text that provides a narrative around the various metrics is a rich source of data in itself – warning of potential delays before other status reporting metrics typically would.

Soft factors – such as political, social and cultural issues evident in the text – are usually disregarded as insignificant, but will often contain the first signs of any upcoming obstacles.

Natural Language Processing

By using a Natural Language Processing (NLP) algorithm, it is possible to extract valuable data from the text, enabling forecasts to be made based on the choice of language in the report. This provides additional insight paths and data points to both inform and improve predictions. In fact, we believe it is one of the strongest indicators of how delivery is actually progressing.

While there is obvious value in these techniques, such analysis was previously not fit for purpose in business applications. Until recent years, machine learning was expensive, demanding resources with high day rates to spend significant amounts of time building and training a model – and requiring computers with high CPU power to process the algorithm.

But, in recent years there’s been a huge growth in interest regarding machine learning. Awareness of algorithms and their benefits has been coupled with a rise in open source software and cheaper CPU power.

Now, automated machine learning tools enable someone with the skill set of a business analyst to train and experiment with various models – without needing any specialist technical knowledge. This opens doors for work such as NLP analysis on status reports; the financial outlay is much lower, and the benefits can be felt across the entire organisation.

Captivating the C-suite

At NTT DATA, we realise this technique can offer more than simply predicting delivery blockers. Undoubtedly, being able to anticipate problems in delivery is useful, however it does not provide the financial benefit that grabs attention at C-suite level. Linking status report findings to benefit realisation analysis makes for a much more compelling case when presented to executives.

If the NLP model discussed above can help anticipate benefit dilution, a business analyst with a few status reports can now offer rapid insights into the likely overall success of a major programme – even while still in mid-delivery.

C-suite stakeholders are not only provided with real-time predictions on benefit realisation, but also with a brand new level of awareness into their organisation’s delivery capability.

At a fraction of the cost, NLP processing of status reports can provide the same depth of understanding and value as a capability maturity assessment. And that, we predict, is worthy of C-suite attention.

 

Are you keeping up with the machine learning and artificial intelligence capabilities that your competitors are employing to take advantage of internal reporting data? Learn more about the potential of machine learning in project, programme management and capability optimisation by contacting Joel.Brocklehurst@nttdata.com


我々の強み

How can we help you

Get in touch