Earlier this month Econsultancy and IBM hosted a roundtable event in Kuala Lumpur where delegates were invited to discuss challenges and opportunities around marketing data.
I’ve already written up the main talking points from the marketing performance management table, and now here is a summary of the conversations that took place on the data-driven marketing and personalisation roundtables.
One of the common themes that emerged involved the thorny issue of ROI.
Malaysian marketers are struggling to prove the ROI case for data-driven marketing, and this lack of budget then holds back any further testing and implementation.
During discussions on the personalisation table there was a feeling that cost optimisation conflicts with personalisation.
How do you ensure that personalisation aids cost optimisation rather than being seen as something that won’t deliver a return on investment?
Delegates with more mature data programmes recommended tackling low-hanging fruit first.
Taking on the easier challenges that ideally require fewer resources can help to prove the business case, get senior buy-in, and also build momentum in the right direction.
This acronym stands for people, processes and technology, and it was a recurring theme across all three roundtables.
Delegates were unsure where to begin with new data-driven processes – do you hire in the talent first and then invest in technology, or the other way around?
Both methods require a high level of investment which is again proving to be a roadblock for implementing data-driven marketing.
This is such a common problem that it’s almost not worth mentioning – we can just assume that everyone struggles with fragmented data sets.
Delegates discussed the various problems that arise as a result of this and how different data can be stitched together to paint an accurate picture of the customer.
This would then allow marketers to truly harness the power of data to better manage and improve the customer journey.
Which data sets are important?
Delegates admitted that they were still having difficulty working out which data was actionable and which could be ignored.
For example, with product recommendations what data should be used to inform which items are displayed?
Do you base it purely on items they previously bought, or a combination of several factors such as demographics and other behavioural data?
This becomes even more difficult when you consider that most companies haven’t stitched together all of their datasets, so advanced personalised marketing in real-time isn’t feasible.
Moving on from test & learn methods
When developing new marketing initiatives delegates admitted they are still using the old method of coming up with a hypothesis and then testing it on a sample.
However this isn’t making the most of the opportunities afforded to us by new data marketing software.
Marketers should really be flipping this around, beginning with the analysis of existing data and trying to make sense of what we already know.
Many businesses still heavily rely on offline methods of data collection, particularly in B2B industries where a lot of customer contact is face-to-face.
This can make data collection more difficult and less accurate because it’s open to human error, which obviously has a negative impact on marketing campaigns.
As one delegate succinctly put it, if you put rubbish in you get rubbish out.
Another popular topic was the collection of ‘soft’ data, such as your customers’ objectives and KPIs rather than personal information.
Marketing messages can then be tailored based on what they are trying to achieve rather than their demographic profile.
This was considered to be a more advanced method of personalisation, but it’s something worth investigating.