To say that we live in an always-connected world is an understatement. We continually consume content, videos, podcasts, apps, and social media on our mobile devices. Even our purchases have become largely online. Despite all this online behaviour, we still live in the physical world. Living in our neighbourhood, commuting to work, doing our favourite activities such as biking, running, skiing or golfing. Whether as young parents, taking kids to daycare, pet owners taking pets to a park, going to recreational facilities or travelling for business or pleasure.
The combination of online and online behavioural data provides valuable insights on consumers for optimizing marketing strategies.
The large scale adoption of IOT (Internet of Things) and the arrival of AR (Augmented Reality) has made the world of offline and online information inter-mingle providing rich view of consumer journeys.
This article discusses these trends going mainstream and the future of marketing strategies when such data is combined with AI.
Identifying the Consumer Journey is still Complex
Marketers are inundated with data points about consumers today. Yet, figuring consumer intent to purchase and effectiveness of specific marketing campaigns has become more and more complex as consumers are utilizing information from multiple channels to make decisions.
15 years ago, a flyer campaign’s success in a neighbourhood could easily measured by the amount of store sales as it is the single driver of sales. But today, a flyer is just one starting touchpoint in consumer journey. The consumer may subsequently search online, look up social media for reviews, do a price check with apps before making a physical or online journey for purchase. Doing attribution with all these touchpoints is complicated without the right offline and online data points.
Offline has Become Online
The Internet of things has become mainstream. We are swimming in devices in the physical world. Many of them capture data (with our opt-in permission) as we live our lives. The best example is location data. It is often tracked by travel, navigation or weather-related mobile apps that record movements and visits of users. Fitness apps or fitness devices (such as smartwatches), also collect this type of data. The main purpose of this data collection is to provide consumer value such as turn by turn navigation or calories burnt etc. This data, with the right privacy controls (when aggregated and anonymized), also provides valuable insights into the physical world behaviour of consumers.
Other offline data collected includes retail purchasing behaviour, at the point of sale using loyalty cards and credit cards. These provide valuable insight into offline retail purchasing behaviour. Bluetooth or wifi beacons installed in stores provide information about time spent in specific isles of retail stores.
Weather is a huge offline overlay on retail behaviour as our daily movements, shopping, lifestyle and even types of shopping are impacted by local weather. Today with Big Data technologies available cheaply, it is possible to take decades of local historical weather information, and point of sales data by SKU, to find correlations. Smart thermostats at homes which reduce energy bills using weather forecast information, are also able to gather home daily and holiday travel patterns of consumers.
From data streaming, low storage cost of data and the ability to analyze and derive insights from large scale data in seconds, it suddenly makes all of the above offline data actionable for marketing in addition to online data.
What Location Data Tells Us
As mentioned in the previous section, location data reveals insights about home, commute and work patterns of users. It also reveals shopping habits and how far consumers take physical journeys to shop. It is a great proxy for lifestyle choices of consumers to understand segments such as pet owners or frequent business travellers or health enthusiasts etc.
The commute is also a great insight to under aggregated insights about users who walk by advertising boards and opens up new information that was never understood in a non-IOT world.
Location data is also a very sensitive area where if the tracking insights are individual and personally identifiable (versus aggregated and anonymized), consumers often tend to disapprove of sharing such data both with big players or simple apps.
What Retail Data can Reveal
Retail point of sale and loyalty card data provides great insight into actual purchasing patterns versus just in-store visits. It is a great proxy to habitual purchases, holiday season and promotional marketing effectiveness. It also gives insights into the physical distance consumers are willing to travel for taking advantage of marketing promotions.
A new wave of technologies that are emerging include smart shopping carts that provide value to consumers such as saving check out times (as items are checked in-out automatically), but can also collect information about their physical retail browsing behaviour and capture intent. Smart AR mirrors that let consumers feel products without physically touching them also can gather intent of shopping choices that consumers are considering.
The Merging Worlds of Marketing
Omnichannel marketing is virtually every brand’s focus today. Traditionally the marketing channels such as Radio, TV, Flyers and direct mail, have typically utilized the neighbourhood or city where consumers live to effectively target them. Marketing channels such as advertising boards (whether on highway or bus shelter or train stops) typically utilize commute and density of traffic of users to identify maximum visibility and effectiveness. For all these types of media marketing, in the past – the main data has been overall population size for campaigns based on where consumers live, work and commute.
Online marketing channels such as online display advertising, SEM (search engine marketing), social media marketing, have typically been more sophisticated (than offline channels) and have been driven by behavioural data such as online profiles of users to identify targeted audience.
However, now with the availability of offline and online behavioural data, the above marketing channels are able to take advantage of insights that were not available before. For example, out of home marketing campaigns are now able to precisely identify microcosms of audiences going by their boards every day using aggregated location data. They are then able to retarget those users on mobile apps online or identify the right boards and times to advertise for certain audiences. Similarly online only marketing campaigns that use search, social or display ads, are able to measure effectiveness of online campaigns not only with online metrics such as click-through, but also offline metrics. They can measure footfall attribution in stores from campaigns using location data or additional sales generated in stores using loyalty card data.
While several data points both offline and online exist about consumers today, the majority of these insights are siloed. For example, the big three (Amazon, Google and Facebook), have a tremendous amount of online purchasing, search intent and social intent behavioural data and hold siloes in their specific specialties.
Retail brands with large loyalty programs often hold siloed insights about online and offline shopping behaviour both habitual and those driven by marketing.
Large publishers and advertising dsps often have siloed insights about online content read, apps/sites visited and location insights about their users that is tremendously valuable for marketing.
Typically all of the above insights are independent of each other and while the big three (Google, Amazon, Facebook), often make it easy for brands to bring other insights to their platforms, they do not allow insights to flow from their platforms to any open ecosystems.
The Coming Together of Trends
In the past couple of years, there has been a growth in the following trends going more mainstream. These trends have been growing as discrete patterns where a combination of offline/online insights are driving marketing. CRM and in-store data is being increasingly used by brands for marketing insights and media spends. The effectiveness of marketing driving store visits and purchases is increasingly being measured and adopted as a key metric and in general advertising exchanges. Thereby allowing tech providers to bridge offline and online identities of users.
In the near future, the growth of these trends will also ensure that insights from each of the circles are linked together, ensuring maximum business intelligence for marketing insights. This will enable more unified decisions based on the true picture of consumers, rather than siloed decision making, for example, looking at just one specific audience parameter (such as behaviour on search engines).
The Key Question of Privacy
As marketing is advancing, connecting dots about users’ online and offline journeys, consumers are increasingly getting worried about their personal data after major scandals in Facebook and other online publishers.
For the premise in the previous section to hold good, it is imperative that consumers willingly participate in the above data ecosystem without fear. This will happen only when the foundational principles of privacy outlined above of consent, control and no-PII are met. For example, the user totally understands any data provided from their end and provides consent, the user has complete control over their choices, and data shared leading to marketing insights end up in aggregate and not use personally identifiable information on habits or behaviour, (unless user provides explicit consent for it).
The Future Tech Stack
Given these trends and multiple providers holding different pieces of the puzzle, the future tech stack of marketing has to evolve and become a lot more open. While players such as Google, Amazon and Facebook will have their own islands due to their individual user scale and data, the rest of the world and retail brands will have to derive their own marketing insights and activate them.
It would involve a data management layer. Data from different sources are assembled in aggregate through APIs and insights can be derived visually from a platform on top. The insights can then lead to the right audiences for marketing, activation of such audiences and measurement of such activation.
As more data on online and offline behaviour is available, it becomes a rich area for AI application. Both supervised learning (predicting which customers are best intended for certain marketing message based on past behaviour), and unsupervised learning (identifying patterns and clusters from online and offline behaviour that can help predict future behaviour), can be applied on such rich data sets. As an additional layer, weather adds an extra dimension on how user behaviour changes with localized weather and can provide rich insights as well. The more various data sources are available (in a privacy safe way) for machine learning, the stronger the output of AI will become.
Over the last few years, marketing has greatly changed due to the vast availability of data. It has become a very focused science that allows retail brands to measure results and to spend very efficiently.
The new trend of offline behaviour of users as a data source, adds a major insight of physical world behaviour. When combined with other online data, it allows retail brands to be smarter in their marketing efforts.
Bala Gopalakrishnan – Chief Data Officer, Pelmorex
Celeste Normington – Head of Sales, Data and Technology Platforms, Pelmorex