What Singapore marketers need to know about making data count

Recognising data as a resource similar to traditional resources like water and energy can give Singapore a competitive advantage and propel the country to become a key player in the global digital economy.

This is integral to Singapore’s pursuit of becoming a leader in fields like analytics and cybersecurity, as shared last September by Chan Chun Sing, Deputy Chairman of the Committee on Future Economy at the 2016 Public Policy Challenge - an annual case competition.

With Singapore’s digital marketing spend projected to reach S$438 million this year, there is no question that data is fuelling our economy’s growth. This is echoed by the recent partnership between Economic Development Board and Google to launch Digitize, a new programme that trains local advertising talent to understand the programmatic technology space.

Yet, it seems that marketers currently do not feel equipped to handle the explosion of data in the ad tech space. According to data from the CMO Council, 30% of marketers feel that they have a good grasp of dealing with data, while 33% reported feeling poor or very poor at analysing data for personalised experiences.

Even with perfect data, campaigns could still be inexact and unpredictable if marketers do not know how to best apply the data. At the end of the day, brands do not need bigger data. They want actionable insights. The real question for marketers is therefore how to mine and aggregate the data to draw useful, meaningful conclusions.

I’m sure most brands have had this line levelled at them: “I can give you mothers in the market for diapers.” In a world with so much data, however, the problem is that none of us really know how this segment was created. What attributes went into figuring out this was a mother and that they were in the market for diapers? Where did this data come from? How fresh is the data?

With no fixed set of standards that defines if the data methodology is really as it seems, it can be difficult for businesses to justify that their data is more accurate compared to someone else who is making the same claim. So here are 3 ways business should manage their data to make them count.

Good data starts with good technology
Marketers have an enormous opportunity to use data to make brand impact amongst its most relevant audiences. This requires a commitment to technology investments, and in most cases, specialised technology tools that remove inefficiencies and produce consistent results.

Marketers can, for instance, replace manual data checking processes by a page-level contextualisation engine that trawls through individual web pages to extract the meaning of the content and append a topic.

There are different arguments for competing methodologies for the categorisation from domain-level to page-level, manual to automated, keyword to contextual – but the key benefit of page-level, automated contextualisation is granularity.

Knowing that a user is interested in a BMW M3 rather than just the automotive industry, for instance, gives valuable interest signals that could be missed without the right technology.

Good data is precise
There are many legacy systems around the audience discovery process, and they often depend on claimed behaviours. As good data is precise, marketers need to be careful and look to more advanced techniques of audience evaluation on real observations.

Creating a useful database is one thing, but ensuring quality is another. One way to suss out the viability of data is by looking at reach and lift. Going back to my previous illustration, lift – the statistical likelihood of a user to act in a certain way or have a certain behaviour in their profile – allows us to say “User A visits the Huggies site to look at diaper products. User A has the most lift to be a mother in the market for diapers.”

Data scientists are then able to look into each data impression to put together an interest-based audience that has similar lift numbers, but on a much bigger reach i.e. how many times User A has gone to the Huggies site. This ultimately provides healthier conversion numbers for marketers.

As part of determining whether dollars are spent effectively, media partners today are increasingly keen to know how accurate their digital campaigns are. There are multiple audience targeting data approaches – from social engagement targeting, which serves ads to consumers who have displayed clear signals on social sites such as ‘liking’, ‘sharing’, or retweeting, to demographic and conversion data targeting.

Essentially, there are infinite potential audience targeting types, but the single most important characteristic of precise data is that it reveals real insights rather than generalised and assumed behaviours.

One campaign executed in Australia is a case in point. An auto company considered as a brand that younger couples had greater affiliation with, was marketed with a lower price point designed to match this demographic.

However upon capturing further intent signals, it was found that an older audience – specifically down-graders and empty-nesters – were also attracted to the brand because of the lower price point. This allowed marketers of the auto company to target a new group of consumers who displayed similar behaviours, but were not initially accounted for.

Good data is reaching the right audience efficiently
With the proliferation of digital viewing platforms, marketers are often pressured to run their campaigns across every channel. This is neither necessary nor the best practice, as data helps marketers target their most relevant audiences strategically.

To begin, marketers should frame their evaluation of each platform, or channel by asking: What is its unique attribute? Which target audience does it serve? Do these attributes align with the business’ needs?

In order to get into the hands of target consumers at the right place and time, many global companies engage real-time optimisation algorithms that help brands look into each campaign impression for best fit, including audience lift or probability for an audience to convert, leading to greater performance and efficiency.

By analysing the highest indexing behaviours exhibited by the existing customer population against interest-based audience segments, the interest and intentions of key online prospects can be pinpointed based on previously collected data. Over time, this results in less impression wastage.

Whilst online ads used to be bought traditionally, like print and TV, we are now also seeing the rise of ads being bought programmatically. Programmatic advertising’s value proposition lies not only in scale, but also its efficiency.

By combining first-party data and programmatic, advertisers are able to curate campaigns for a really targeted audience, and achieve their media planning objectives in an efficient and focussed manner.

Less money spent, and less annoyed consumers. Which digital marketer wouldn’t love that?

Whilst there are limitations to the above measures due to the reliance on certain networks and tools, the beauty of digital is that on a broader level, it can be measured in terms of outcome, efficiency, and transparency.

In an increasingly competitive digital space with heightened consumer expectations for relevancy, businesses armed with critical insights from the aggregation and analysis of data coupled with speed to market will continue to take the lead in the marketplace.