However, the report – which features new interviews from senior practitioners, combined with insight from related Econsultany surveys – also suggests there are big challenges still to overcome.
The journey from a descriptive to prescriptive model
First, let’s remind ourselves how organisations were engaging with predictive analytics back in 2016.
The first report outlined three different models, including descriptive analytics, the first and most basic stage, which involves using raw data to summarise historical patterns and trends.
The second is predictive analytics, which attempts to use that historical data to tell the organisation what will happen if it takes certain action.
Lastly, there is prescriptive analytics, which aims to thoroughly analyse an individual’s behaviour, and recommend the best next action that will deliver value both to that individual and to the business.
Ultimately, it was revealed that the majority of respondents were engaged in some sort of predictive modelling, but more than a third were using descriptive analytics only.
So, has this since changed? Econsultancy’s latest report focuses on four key areas of focus for organisations still on their way to predictive maturity.
Silos still plaguing organisations
One of the biggest changes from the previous report appears to be organisational attitudes towards data and analytics. Instead of striving for the right type of data, organisations are now moving towards utilising the data they already have to help drive business growth.
In other words, organisations are now being led by data rather than merely collecting it, moving on from descriptive analytics to more mature models.
Despite greater ambition, and a shift towards a data-driven culture, siloed data (and attitudes) still remain a big problem.
This is largely due to the push for data stemming from either the top or bottom of organisations, leading to a lack of willingness to share data, as well as confusion over where analytics should sit within the company.
The solution to this for a lot of large organisations could be to recruit and retain data specialists, with the end-goal of keeping analytics independent from IT and marketing. Essentially, to establish an in-house agency of sorts.
Naturally, this poses many challenges, with two of the biggest being finding resources and keeping up with technological change.
Expertise in data management
While culture is vital for organisations making good use of data, it is of course just one part of the story. More important is the data itself – i.e. what data organisations are collecting, and how they are managing and integrating it.
Interestingly, despite ‘big data’ being somewhat of a buzzword for a number of years, and its importance being recognised by marketers, the majority of organisations do not see themselves as experts when it comes to data management.
The 2018 Digital Trends report found that just 19% of respondents strongly agree that they have some access and control over customer and marketing application data. What’s more, 39% of respondents rated having access and control of data as being a ‘difficult’ part of customer experience delivery to master.
Again, with data typically being collected by different teams and in different formats, barriers to its integration tend to be organisational and cultural rather than technological.
This tells us that organisations should indeed be focusing more on demonstrating the value of existing data rather than striving to find the ‘perfect’ subset.
This can also help to resolve some of the aforementioned problems around data sharing, with senior staff being more likely to put resources behind data integration if they are aware of the benefits.
Turning insight into action
According to Econsultancy’s Dark Data report, released last year, 40% of the businesses surveyed rated their ability to act on insights derived from customer data as ‘poor’ or ‘very poor’.
The situation has been improving, however, as those describing it as ‘excellent’ rose from 3% in 2016 to 11% in 2017.
One of the main barriers to implementing personalisation is the amount of content it requires. This means that, to a certain extent, automation can be a solution, helping to free marketers from the exhaustive task of testing multiple variables.
Meanwhile, with analytical teams now ‘top and tailing’ marketing – i.e. providing insight needed for campaigns, as well as analysing results – ensuring that insights flow into action can be tricky.
This means that, while creating separate data teams might help solve problems in recruiting and retaining specialists, it could also potentially lead to less integration with the wider business – something that individual organisations should take heed of when deciding on structural change.
In conclusion…
While it’s good news that attitudes towards predictive marketing are shifting, the quality of data, the ability of marketers to collate it, and the skills required to analyse and integrate it are still proving to be big obstacles.
The good news is that, as machine learning in marketing continues to evolve, greater personalisation at scale is becoming easier to implement. What we cannot forget, of course, is the age-old caveat that the insights delivered by a machine will only ever be as good as the data that goes into it.
As a result, real predictive maturity will always rely on us mere mortals, and the extent to which organisations are investing in the skills and internal structure required to reach it.
Related articles:
Comments