Adverai applies machine learning and neural networks to make sense of all marketing tools, platforms, and datasets. We spot patterns in billions of data points to help brands deliver better marketing – at scale.
Making sense of billions of data points requires a radical overhaul of the way in which we handle marketing data. Adverai’s cloud-native platform MetaOS works by transforming massive datasets into accurate predictions in order to provide you with explainable optimisation advice that helps you increase your sales, fine-tune your advertising, and make your marketing strategies more successful.
Connect the marketing data ecosystem
The first step we take in understanding marketing effectiveness is to connect the entire marketing data ecosystem for your brand. Adverai automatically replicates data for every dataset and source connected to your marketing. We replicate data from digital marketing platforms, cinema, tv, and out of home advertising to external market condition data such as weather, economic indicators, and external events. With our operational data platform, you can bring everything together in one place, at scale and at speed.
Automatically organising data from every marketing channel in one place allows you to track effect and monitor spend for every campaign and ad in real-time. More importantly, it creates the foundation for a system of intelligence across your diverse systems and datasets.
Time as the primary axis
From inception, we made the decision to engineer our data platform as a time-series platform. We quickly realised that we needed to not only observe the current state of marketing but to measure how marketing and consumer response changes over time. We rely heavily on time to preserve information about change.
By centering our platform around “change”, we can start to identify the interesting patterns that affect marketing in real-time.
A time series is a series of data points indexed in time order. Most commonly, a time series is a sequence taken at successive equally spaced points in time; a sequence of discrete-time data. Examples of time series datasets are counts of sunspots and the daily closing value of the Dow Jones Industrial Average. Time series are used in statistics, signal processing, pattern recognition, econometrics, mathematical finance, weather forecasting, earthquake prediction, electroencephalography, astronomy, and in any domain of applied science and engineering which involves temporal measurements.
Time-series data collectively represents how a system, process, or behaviour changes over time and where time isn’t just a metric, but a primary axis.
Question and answer sets
Within the billions of datapoints we collect and the 100s of patterns we detect are a myriad of questions waiting to be answered.
Ask Meta is used to obtain immediate answers to the big questions that you have as an advertiser, it detects patterns in massive marketing datasets and allows you to ask any of 100s of defined questions with instant answers that are explained using clear and insightful visual narratives.
We wanted to build the ultimate question and answer system for every marketing channel and platform to give marketing teams the tools to make better informed real-time decisions.
Meta allows you to explore your marketing data in a simple and beautiful visual way, and find sides to the story that you never imagined.
Making sense of consumer responses
To makes sense of fast-changing and diverse consumer responses we need to find new ways to build machines that predict human behaviour and build social intelligence.
Inside Meta is the ability to understand the emotional responses to campaigns and ads, to track the distribution of comments, likes, and reactions for each channel, and to track the emotional engagement surrounding posts. By comparing the volume of emotional reactions to your campaign effects and understanding the connection between creativity and emotional engagement we can start to understand what works in advertising and why.
Plotting the outcome of what-if scenarios and finding new ways to optimise and automate marketing requires a state-of-the-art approach to data replication and data storage. We have built a fabric of neural network architectures and models, combined with classical statistical techniques, with a clear path to generating causal inference.
Adverai’s goal is to move from mere correlation and pattern spotting to a true understanding of cause and effect.
Adverai’s state-of-the-art machine learning platform, with pre-trained models and customer-centric tailored models has blazing fast training performance and high accuracy. Our machine learning serving gives unmatched scale and speed required for advertising intelligence.
Neural Networks and Model Zoos
MetaOS’s advanced pattern spotting capabilities enable you to create intelligent time-series for all of your marketing interactions.
Adverai’s technology makes innovative use of machine learning and deep neural networks. We create models that take metrics, consumer responses, and creative attributes to predict how a consumer will interact with ads or select the ad that will result in the highest increase in engagement over time.
We provide the data platform that gives you with the complete understanding of your marketing data ecosystem.
Detecting and classifying significant events.
We believe that it is important to keep you informed and help your marketing teams stay ahead of the game. We keep you informed about the significant events and changes that surround your ads and campaigns and provide clear optimisation advice about the correcting actions you should take.
We classify our events as milestones, anomalies, spikes and drops, or warnings.
The four event types in Adverai Meta.
Milestones: When a campaign, activity, or channel approaches, reaches or crosses a milestone.
Spikes and Drops: When a campaign, ad, or channel drops or spikes significantly over the last hour, day, week, or month
Warnings: When a campaign or ad performance is poor and requires immediate attention.
Anomalies: When a behaviour is unusual about campaigns, ads, brands, or channels across time.
Designing a feature rich and frictionless user experience
At Adverai we focus on building tools that are frictionless in their use and that help you forecast, optimise, and automate your marketing without relying on an army of data scientists.
Meta is your personal data scientist, with sophisticated features that can detect unusual events, recommend how you can optimise your spend, and model the outcome of what-if scenarios, all supported by evidence-based narratives and clear visual explanations.
Building the automation machine for advertisers
Recently, my colleague Mathias wrote an excellent article about “The future of work for marketing” where he explores what we can we expect from our day to day work in marketing when automation surges and some of our best-performing colleagues are machines.
Automation is the technology by which a process or procedure is performed without human assistance. We believe that machine assistance will help us keep pace with today’s rate of change, and that automation machines to help us deliver it.
“AI is enabling the automation of an increasing range of functions, which until recently could only be carried out by humans”
Building meta-understanding and automation machines for marketing requires a state-of-the-art approach to data ingestion, processing, and analysis using deep neural networks, statistical computing, and causal reasoning. As an applied AI company building the next generation data, computation, and experience engines for marketing we seek out and solve these hard problems.