Will marketing become more human thanks to technology?
Let’s ask the question again: will marketing become more human thanks to technology? For the B2B sector, the answer is yes, according to Forrester. For a few analysts at the global market research company, in 2021 technology will help enhance marketers’ and sellers’ human touch. “COVID-19 has changed B2B marketing and sales - perhaps forever. Savvy marketers and sellers will seize the moment and transform their approaches to become truly buyer-centric”.
From an agency/consultancy and a tech vendor perspective there are a few questions to consider, when is the right time to suggest to a client to use human and/or technological approaches? When to advocate for one and/or another? How scalable is a human-led analysis in comparison to platforms that combine AI, natural language processing, machine learning and linguistic rules?
To delve into these and a few other questions we had a chat with two different players in the industry:
- Oliver Lewis, global head of insights at Convosphere, a London-based business intelligence and insights agency
- Ben Sigerson, VP of Solutions at Converseon, a platform that combines machine learning with social and voice-of-customer data
Oliver agrees that tech will make the marketing and sales process appear more human, considering how tech enables digital content to be served to us as consumers based on the signals we give off every time we “behave” online. “These methods can be applied at scale to the B2B process and help sales and marketing professionals conduct their jobs at a previously unattainable scale,” it will help with decision making.
“I think there’s definitely something to the point that tech (or more specifically, AI) is enabling people to spend more of their time on the ‘human side’ of their work (the stuff they’re best at) and less of their time on rote tasks,” in Ben’s opinion. He argues that in the case of social and consumer intelligence, also true for AI enabling insights professionals and marketers to focus much more of their energies on data interpretation and analysis.
“That means that analysts can actually spend their time analysing - assigning meaning to metrics and translating data into a set of actions/implications for their business - rather than performing repetitive tasks like data cleaning and QA,” Ben says.
Eliminating the need for manual analysis
A few platforms mentioned in Forrester’s AI-Based Text Analytics Platforms report suggest they eliminate the need for manual analysis to uncover trends and opportunities for brands. For Oliver, these platforms may well eliminate the need for extensive, labour intensive and, frankly painful, manual analysis processes (extensive coding frames, regular checking, etc.), but, as he states, humans are required to ensure the trends unearthed are salient and the opportunities are grounded in the real world.
Ben argues that the act of interpreting a trend or theme in your data is still largely human-led. “But increasingly, AI-powered text analytics enables human analysts to discover those trends or themes much faster than before.”
Human-led analysis & AI/ML are complementary
We have asked Oliver and Ben how scalable a human-led social data analysis is in comparison to platforms that combine AI, natural language processing, machine learning and linguistic rules.
Ben helps us understand that we shouldn’t put human-led analysis and AI/ML at odds with one another. “Rather, I’d position them as complementary: AI-powered NLP, when built the right way, enables a far more efficient human-led approach to social data analysis,” he states saying that Converseon’s autoML platform, Conversus, actually places an emphasis on “human-in-the-loop” machine learning.
For Oliver, all the technologies mentioned above are of great help to Convosphere’s analysts, “assuming the technologies are equally adept at multiple languages (this is most certainly not always the case)”. “Few people would disagree that they offer scale, but scale is not always crucial.”
The global head of insights at Convosphere also explains that not all projects require exhaustive analysis of a colossal data universe. “Sometimes we are looking at very niche topics where big data does not exist. Sometimes we do not need to consider ALL available data where a sample will suffice,” he comments.
Ben adds that social and voice-of-consumer datasets are characteristically unruly and large; they contain incredibly valuable insights, but those insights are often “diamonds in the rough” that need to be mined in some way.
“When mining for those insights, there has historically been a difficult tradeoff between accuracy and scale. I could have humans perform the analysis very accurately, but this approach could not scale to large sets of data. On the other hand, I could use automated methods to analyse this data at vast scale, but the technologies traditionally employed to do this have been woefully inaccurate. Current AI-based approaches help to break this tradeoff,” in the opinion of Converseon’s VP of Solutions.
The need of people to understand cultural insights
Oliver mentions the “blind alleys” that technologically driven results can sometimes send us down. “These can sometimes weigh heavily on the analysts’ time, negating much of the efficiencies originally brought about by the deployment of the tech.” In his opinion, however, 100% of the time we need people to understand the broader cultural, societal and
political forces at play and interpret, explain and translate the findings.
“Our role is to help our clients make informed business decisions through the analysis and interpretation of social data. We need skilled human researchers to do the interpretation of this data. Their job is to craft a story from the various data points we work with. Technology is there to support this ultimate endeavour, not lead it,” Oliver says.
On the other hand, Ben explains that AI-powered natural language processing is at the heart of what Converseon does: “it’s the technology we use to provide enhanced social and consumer intelligence to our customers.” “By using AI-powered NLP to analyse social and VoC datasets, we can not only analyse at ‘big data scale’, but we can also approach, and at times surpass, human-level accuracy in our analysis,” Ben states.