2019 Ambitions & 2018 Successes: Converseon’s year in review
We wish for all our friends, clients and partners a happy, healthy and successful 2019. This year marks 11 years of immersion and innovation for us in machine learning based social listening and insights, and we fully expect 2019 to be the most exciting yet.
To those who know Converseon, we continue to expand and adapt our technologies and services to the emerging environment. We joke only half-heartedly sometimes that we are a 15-year-old startup. Our original birth as one of the very first pure play ‘social agencies” (at a time when “social media” was not a common term) led to being named Digiday’s Top Social Agency. We evolved from that phase to concentrate on our social listening technologies in 2008 and was named Forrester Leader in Enterprise Social Listening in 2010, and Strong Performer in 2012/14. At that point, we did a slight pivot again in 2014-15 to focus specifically on our AI technology for processing and understanding the massive reservoir of consumer text, specifically in social and related voice of customer areas, where we were finding very high client demand. It was a move we certainly do not regret.
2018 was an exciting year for us as we have forged ahead again to be a front-runner in our category by helping our clients accelerate insights and drive business advantage by accurately and effectively mining this massive, complex unstructured consumer text data at human level precision at the speed and scale that only software can provide. For many organizations, this unstructured content is poised to become 90% of all their stored data. Yet most companies are processing less than 20 percent of it so far (according to Forrester Research), missing a powerful source of competitive advantage.
At the dawn of 2019, we are pleased to be positioned as a clear leading solution for all companies who want to get better insights, action, and value from this data. And while 2018 was a year of rapid evolution here, 2019 is poised for even greater growth and success for us, our partners and clients. We not only provide custom/advanced language models (both prebuilt as a library for instant use and built to spec), we also are enabling our clients to rapidly build their own models through our “no code” machine learning as a service platform, Conversus.AI. In short, we provide the tools, platforms, and services needed to translate complex consumer and social text into meaningful and actionable insight at massive scale and with high accuracy.
Here are some key highlights:
- We had strong revenue growth in 2018 and achieved more than 95% client renewal, validating the value our AI-powered text analytics technology and associated programmatic insights are generating.
- Clients include some of the most respected companies in the industry, including IBM, Dell, Uber, Walmart, among many others
- Our revenue per client has increased by approximately 80% as our models drive greater value and adoption in client organizations
- In mid-2018 we took our first external Series A investment and successfully applied it to accelerate evolution and scale out of our technologies. Even with the investment, we achieved profitability again this year.
- We were the first to launch a comprehensive library of pre-built machine learning models by industry and use (customer experience, customer care, brand tracking, etc.) for immediate subscription, use, and deployment. This is critical to accelerating the use of AI and reducing costs to clients.
- We have transitioned from reactive, descriptive analysis to predictive and prescriptive “on demand” analysis/insights via our highly granular and accurate data and machine learning technologies. We have demonstrated clearly that our machine learning powered data can predictive key business outcomes, including sales and the results of survey-based brand equity trackers, to just single out a few areas. Predictive social is an area of clear focus in 2019. See how presentation redid on how to predict the future with social data for a recap.
- We have and will continue to build out other bespoke models for our clients that can be immediately deployed programmatically with a wide range of key social listening/management and business intelligence platforms. These include Brandwatch (the largest social listening solution after its merger with Crimson Hexagon), Sprinklr, and others, along with BI platforms like Tableau and Domo. Our models are designed for more sophisticated users who demand high data quality and the ability to use machine learning to classify language “like humans do.” Our partnerships are flourishing and we look forward to even greater focus here moving forward. Ecosystems matter!
- We are and will remain at the forefront on model and data transparency, providing clear F1 and other performance scores with each model prior to deployment. We applaud industry efforts in these areas including the IAB Data Transparency Initiative and have advocated for similar labeling requirements for social data. advocated for similar labeling requirements for social data. We hope to work with the industry more generally to move this issue forward and hope all clients will also demand this level of transparency and data quality to ensure the elimination of unintended bias. What we do is too important to not get it right.
- We have expanded our models and support to more than 14 languages and continue to expand to others based on client needs.
- Beyond our model building and insights, our Conversus.AI machine learning as a service platform now puts the immense power of AI directly into our client’s hands, including subject matter experts who don’t have data science skills, to better take control of their data and application. This “no code” approach to machine learning puts human experts in the loop directly and is highly transformative by empowering our algorithms to directly learn from these experts and then project and apply that “intelligence” broadly to the enterprise. Combining the best of human expertise and AI is a very exciting era of growth and we applaud the “no code” revolution and glad to be a part of it along with some others forging the way ahead. In 2019, we anticipate a wider range of client experts involved in building models simply and easily in areas ranging from product development, to brand management, crisis and more. Better yet, these models can be built and modified to meet the specific taxonomies and perspectives of clients. As Mark Andreesen has said, “if software eats the world, models will run it.” We are that model factory.
- We have evolved far beyond just sentiment and emotion, into areas like “trust/distrust,” human motives, advocacy and more as we find and unearth those insights in human language that can help change business and drive competitive advantage. And we’re making these easily and effectively available to use as needed and in real time.
- We are also expanding the applications of our models from social primarily to other sources including call center transcripts, long-form survey verbatims and more. We seek to help companies understand their true voice of customer, from all sources, with one clear and accurate “truth.”
- We continue to evolve our solutions in 2019 to stay ahead of the game – to deliver better, faster and cheaper. Our models now outperform individual humans in most cases and we will be applying even more technology and techniques this year especially in the areas of transfer learning, adversarial model building and other approaches built on the tenets of openness, transparency, accuracy, fairness, and impact.
We want to of course thank all our clients, teammates and partners who continue to challenge us to grow and innovate on an ongoing basis. Their challenges make us better and allow us to build on our rich heritage of innovation and excellence as leaders in social and VoC intelligence. We can’t wait to get the new year underway and hope that we have a chance to work with all of you to further the evolution and value of the industry.
Perhaps most importantly, we wish you all a happy and prosperous New Year.