Can marketing survive without humans?

09 Sep 2015

You’re at the checkout but nobody’s there. You wait, tapping the ball of your right foot, hoping the patient, well dressed lady behind you will empathise. Ahead of their Man vs Machine webinar on 23 September, Emarsys look at automation.

A large, shiny sign with an arrow catches your eye. You’re drawn to it, guided to a strange talking computer with a mysterious lady’s voice. You’re firmly but fairly told to scan your items, put them into the bag on the right and hand over your card details to the effeminate bot. You comply. You’ve come in and out of the store without seeing a soul, other than the lady who was behind you, who has become less patient since being shortchanged, and there’s nothing she can do about it.

The truth about humans and technology.

Technology is replacing human jobs everywhere in the Western world: passport control, superstores and more recently even surgeons have been effected. Technology that replaces human jobs is something we’ve complacently come to accept. The population is growing, and technology is taking over to keep corporate budgets down, removing the need for humans to work. With so many people in the world, it’s impossible to remember your individual customers by name, or even meet them in the first place. This made learning customer likes and preferences a true challenge, something had to be done.

What’s all this about automated content?

For more than two years, 79% of top-performing companies have been using marketing automation.1 On top of this, almost a fifth of companies who use it find their revenue increases by 75% or more.2 Marketing Automation is delivering results, and with global marketing spend on $22.6 billionâ—âset to increase by 50% by 2018,3 it’s not going anywhere. Per dollar spent, content marketing produces three times the leads of paid search4 and the number of searches for the term “Content Marketing” has increased by 400% since four years ago.5 Content is evidently becoming even more important, but bearing in mind the latest technology, does the future of content still involve humans?

Semantic algorithms already exist to create and evaluate content for marketers. What’s more, they’ve even been proven to produce more revenue on their own than with human input. Natural language firm, Persado, created an algorithm to determine subject lines for emails. When they’re used, they can get double the open-rates of hand-written ones.6 The methodology is relatively simple: the programmer scores the effectiveness of all possible words in a database. The algorithm reads the database and analyzes all variations of a message. It selects the most effective one, based on the word scoring, and writes a subject line with a known emotional pull.6

Automated content technology can do much more than subject lines.

American-based technology company, Automated Insights, has created Wordsmith. This tool takes data from any source and writes about it in plain English.7 Wordsmith can make near real-time commentary of sports games (it can even refer back to historical sporting events like human commentators do to give further context for the audience). Furthermore, it can write detailed reports and summarize complex data in a simple way, in any tone of voice. Enterprise companies such as Samsung, Associated Press and Yahoo are all already using Wordsmith.7

Even a few years ago, Nimble Books programmed an algorithm to write entire books on any topic. According to Forbes, ‘it gathers articles…brings them together, organizes them, generates a table of contents, and voilaâ—âit’s a “book”’. In 2008, a robot combined the plots of Leo Tolstoy and Haruki Murakami.8 It’s a novel idea but the definition of a “book” here should be considered, perhaps it’s more of a sophisticated amalgamation of text. According to Aberdeen Group, Sales and Marketing both rate lead quality over quantity to build higher revenue.9 Meanwhile the Lenskold and Pedowitz Groups found that 68% of successful marketers link good quality leads to high content engagement.

Automated content may be more optimized, and may achieve higher open or click rates, but the writing itself will likely not be of the same quality. If this was the case, content marketing would likely lose a significant amount of value to customers.

Machine Learning.

Languages are fundamentally formulaic: they follow distinct sets of rules based on spelling and grammar. Once a software is connected to these rules, they can be broken down and “taught” to it. Theoretically, computers could write any type of content: the only missing item would be a human understanding of tone and context. According to The Guardian, content requires “original thinking to come up with efficient ways of solving business problems, usually requiring a lot of human empathy along the way”.10 So if a brand’s community is searching for original and innovative content, intertwined with human industry experience, they won’t find this via artificial intelligence. Unless they can educate the machine.

Amazon have come up with something called Amazon Mechanical Turk, which has the tagline “artificial artificial intelligence”. This is a website on which any human can access a database of Human Intelligence Tasks (HITs), or tasks which require human, as opposed to artificial intelligence.11 Tasks are low paid but you can do a lot of them very quickly, like categorizing the tone of an article or rating the sentiment of a particular webpage.11 At first glance, this artificial artificial intelligence looks like a campaign against artificial intelligence: perhaps a helping hand so humans make money using skills only they, and not technology, have. On the other hand, if thousands of humans sign up to perform the same HITs, Amazon would build a rich database, the human tone and context that technology currently lacks. Ideal for perfectly automated content in the future.

Tech that supports (and doesn’t rival) humans.

Machine learning has turned strongly in the direction of the commerce industry. As 74% of customers get frustrated when website content appears that has nothing to do with their interests,12 programmers found that customers need personalization, and businesses are willing to pay for it. A good example of this is Emarsys Smart Insight: this customer intelligence technology finds shopping behavioral patterns between customers and uses them to create ‘personas’ for brands.13 In this context, machine learning allows each persona in your marketing database to be attributed different content and marketing designs. This makes each campaign more relevant to the person receiving itâ—âwith this technology, a human creative flair is certainly still required to write and design various content.

Amazon began the trend of predictive product recommendations some years ago. This technology “knew” the products customers would like to buy often before they knew it themselves, and today it’s responsible for 35% of Amazon’s entire revenue.14 Predictive product recommendations observe customer behavior to collate customers into ‘personas’, groups of people who have similar patterns of customer behavior. At the time, this technology was only available to enterprise companies with extremely large marketing budgets. Today, this technology is known to increase revenue from marketing 6 fold, yet 70% of brands still do not personalize.15

So can marketing survive without humans?

Marketing works when it genuinely educates, entices or advises its audience with accurate and (seemingly) impartial content. It draws on research, user experiences and product specifications to persuade its prospects. Ultimately, technology is only as strong as the data fed into it. To achieve a persuasive, interactive and convincing relationship with customers (and without marketers), machine learning would have to be more advanced than it currently is.

Conclusively, the ideal of automated content would be attractive for a business in terms of overheads, much like self checkouts in superstores. While it’s not 100% possible yet, automated marketing could already replace roles of any human marketing team. Commerce marketers are patient, often well dressed people. But it’s important for them to keep their fingers on the pulse of their industry. It’s no good waiting in line to for human interaction, hoping they don’t get shortchanged, as by then, there’d be nothing they could do about it.

Sources:

1. Email Monday

2. Hubspot

3. Chiefmartech

4. Kapost

5. Captora

6. Fortune

7. Automated Insights

8. Forbes

9. Aberdeen Group

10. The Guardian

11. Mturk