All marketers know how important words can be, especially in communicating brand messages and connecting with customers. At last week’s Big Bang 2019 event held at the Science Museum in partnership with The Drum, Marin Software’s head of marketing Wesley MacLaggan, illustrated how to get more out of your paid media strategy panel.
Reliable translations with Google
Referencing Google’s recent Super Bowl offering, MacLaggan he showcased how Google Translate service pointed out that although “words can hurt and sometimes divide,” the most translated words in the world are “How are you,” ″Thank you” and “I love you.”
What was previously a chunky piece of tech that translated texts too literally, now serves as a useful and essential piece of kit used many people around the world to translate more than one billion words every day.
The Google Translate app is an easy-to-use tool, that translates text in real time when users point and aim their camera lens at a piece of foreign signage. “Using AR, it matches what you're seeing,” said MacLaggan. “It matches the font and the colour, as if you're looking at the product in a language that you can understand.”
MacLaggan cites this piece of technology as a reimagining of how we ought to work.
“The Google team could have reinvented a dictionary but they wanted to make it better; to make the app a real experience,” he said.
Previously, the app translated sentences by looking at individual words and using a statistically-derived dictionary to transform them, upholding rules to manage conjugations and checking that word endings correspond. However, this phrase-based approach often resulted in a random assortment of words that lacked the natural fluidity of speech. Which is why the Google Translate team took it upon themselves to tackle this issue, fuelled by the aim ‘to make computers smart and make lives better.’
By building a Google neural translation system and crafting an artificial neural network – based on data provided by New York and Brussels, who work with the UN and EU to translate documents into multiple languages – the translation system was able to significantly improve.
"However great the Google neural network technology is, unless it’s got the data on which to learn, it’s not going to get the results," he added.
Keeping the customer in mind
Of course, the end result is what matters most to users. MacLaggan reminded marketers that they need to take advantage of the information they receive along the customer journey as it can prove useful in answering essential questions about their target audience. But accessing this data in its entirety can prove difficult.
To begin with, the consumer journey is fragmented. Customers receive ads across different platforms and may make several searches on various ecommerce sites before actually making their final purchase. “If you're looking through the lens of only one of those publishers you won't have the complete story,” said MacLaggan.
He recommends looking at the consumer’s journey more holistically. “There’s a whole world of information outside of the actual auction and customer journey that can really influence the performance of your campaign,” he said. “Firstly, use all the data available to find out who the customer is, how much they're worth and then you can establish your relationship with them. But also consider what’s going on within your promotional calendar; perhaps an event – like a sale – has inspired someone to come back and complete the deal. All of this is data that can be fed into the machine learning systems to optimize your digital campaigns and help to deliver better results.”
How to stand out from the crowd
MacLaggan suggests that marketers focus on the following three aspects of data if they want to stand out in the competitive market space.
- Audience: By considering the lifetime value of prospective customers, marketers can work out their specific needs and reduce overpaying or underpaying for certain consumers. “Figure out and understand what’s driving their decision,” said MacLaggan. “If you look at this on a geo-basis, there's probably going to be clusters around university towns where you see very different behaviours from more typical neighbourhoods. This is the sort of understanding you need from your audience.” Leveraging data and reading cross channel signals can drive campaigns by as much as 30-40%. Capture the full range of interactions from your target audience; purchase cycles can take longer than expected to convert customers. “Our goal is to bring together all of those touch points,” said MacLaggan. “So, you can be reactive to changing marketing conditions with those top-of-funnel signals and understand the value that the different campaigns are driving.”
- Product information: Dynamic campaigns and ad formats dominate ecommerce sites and social media platforms, which mostly use feed-based ads with the information being sent back to the publisher. But there are ways around this. “Think about product sales across channels; it’s another opportunity to leverage information,” said MacLaggan “If you’re selling on Google and Amazon, compare how products are doing across these channels. Figure out what’s working and what’s not; it may help to identify opportunities for driving better performance across channels. They may be the same but relatively-speaking, if something is top of the rank on Amazon and bottom of the rung on Google, there’s probably something going on there.”
- Context: Understanding the circumstances around a customer’s journey will also be advantageous. Marketers should programme their promotional calendars into their machine learning to ensure that it can factor in any activity around these events. "Have machines process your promotional calendar – to determine what will happen before during and after around specific events,” MacLaggan said. “That way, it can provide tailored results.”
By giving machines more tools and data to work with, marketers can drive incremental performance relative to their activity. MacLaggan suggests investing in Marin’s newly-released Search, Social and Ecommerce guidebook which provides more practices on driving business growth and how to maximize the performance of machine learning. He urged marketers to rethink their relationship with customers by welcoming in the right data.
“There are many possible futures; the world as it is, it’s not a given. We can reinvent it,” MacLaggan borrowed this quote from tech legend, Tim O’Reilly to conclude the session and remind attendees that they’re able to shape the future direction of the industry.