Many business owners and department heads have heard of Artificial Intelligence, “Machine Learning” and predictive models, but aren’t quite sure how these technologies could apply to their business. Today there are solutions that let you know, for example, which customers will stop paying, which will cancel or which will see a significant drop in revenue, and they aren’t expensive services but accessible to every company. Using them in the coming months will largely determine the ability to emerge from the recession and be efficient and competitive in the market.
Descriptive analytics vs. predictive analytics
Most companies’ information systems rely on traditional reports from their ERP and CRM management applications, or even Business Intelligence solutions.
With them, they can get reports and dashboards with historical data to analyze company performance and trends across different business areas.
These tools are great and have been essential for many years, but they only provide information about the past. Today you can take data analysis a step further and get predictions about the future through customized predictive models.
Being able to make decisions knowing what will happen in the future —which products we’ll sell, what the repeat rate will be, which customers will default, how much my customers’ revenue will fall…— provides key insight for running the company more rigorously, especially in difficult times like the ones we’re living through.
Practical use cases for Artificial Intelligence
In every crisis, most companies in every sector will suffer an economic impact that mainly shows up as revenue drops and customer loss.
For both, there are solutions based on predictive models that can significantly help minimize the impact.
Revenue drops and customer loss
We’ve found this is a common concern for every company, and introducing artificial intelligence can mitigate it significantly.
These predictive models forecast the churn risk of customers / subscribers / students / users / members… depending on the sector.
The risk can simply be “the probability that a customer will cancel” in subscription business models, or risk can be defined as potential customer “revenue drops”:
- For sectors such as Newspapers/Education/Insurance/Professional Associations/… and any subscription-based business, the model provides a monthly prediction of each customer’s/subscriber’s churn risk.
- To cover revenue drops, the model is fully cross-industry and predicts customers’ revenue drops over a defined period (for example, 1 year), classifying them into 4 segments:
-
- High risk – Customers forecast to see a significant revenue drop
- Moderate risk – Slight revenue drops
- Stable – Those whose revenue would stay as it is
- Growth – Customers expected to increase revenue
-
The model shows the average revenue loss for each segment over a 1-year horizon, as well as the individual values for each customer.
With these two models, companies gain very powerful tools to face the crisis and the subsequent recovery, and can launch and focus the most appropriate sales, financial and marketing activities and decisions from minute one.
The two points above are just clear examples that address very common concerns right now, but there are many other business areas that can be optimized with predictive models. Here are some examples by sector:
- Education – Improve conversion rates from interested students to enrollments
- Real estate and automotive – Optimize lead conversion from first contact to visit or final sale
- Staffing agencies – Screen and optimize candidate search and selection
- eCommerce – Improve repeat purchases after the first order, and customer value and potential models
- Leisure, tourism, travel, portals, call centers, energy, telcos – Run more effective sales campaigns, improving sales conversion rates in a market where everyone will be rushing to capture leads
The democratization of these solutions makes them accessible to all kinds of companies, since they are subscription services (SaaS) at very reasonable prices, with implementation projects taking just a few weeks.
Key concepts for economic recovery: efficiency and adaptation
To paraphrase Darwin, we can say that in the current economic climate “The companies that survive will not be the biggest or the most powerful, but those best able to adapt to the circumstances of the pandemic”
Relying on data for decision-making is increasingly the way to improve efficiency and adapt your eCommerce business to change. Being able to adapt quickly and know how actual activity will perform through predictive models is a tool with extraordinary potential for tackling the recovery.
Running a company on historical data that has lost all its value and relevance is not the same as relying on constantly updated information, based on machine learning technologies that learn on their own and quickly adapt to circumstances that are more changeable today than ever.
To survive and accelerate, it’s crucial to know what will actually happen (with its variations and alternatives), not what happened before or what procedure seems to dictate should be done.
With this privileged information we can steer the company toward its goals and base our decisions on science rather than gut feelings, opinions or past data.
It’s in this context that the two concepts mentioned at the start, efficiency and adaptation, matter more than ever.
- We’ll need to be efficient to accelerate, and have “technological weapons” to achieve more and better at lower cost
- AI and Machine Learning tools ensure much faster adaptation to changes in customer behavior
Acting with urgency is also important, because the democratization and cost of these solutions make them accessible to everyone, and falling into “analysis paralysis” will only leave you less prepared and competing at a disadvantage.
