Rainmaker: Early Bird Registration Ends April 30th, 2024 | Reserve Your Spot

Contact Sales
3D illustration Rendering of binary code pattern

Optimizing Revenue with Artificial Intelligence

August 10, 2020

By Chanan Greenberg, SVP & GM, High Tech Products

Artificial Intelligence is here now. It is on the budget of virtually every CIO of every major corporation and it is cited as the number one area of investment of 300 plus executives to help manage and improve revenue in the recent State of Revenue 2020 Report. AI is showing great promise helping optimize deals and in the short term will become a competitive differentiator. This blog will focus on a several key questions including:

  • AI – Why now?
  • What is an effective approach to implementing AI?
  • How can AI help companies optimize revenue?

AI – Why now?

AI is not new. As a concept it was contemplated by the ancient Greeks. As a practical pursuit it has become an area of research and investment spanning more than 6 decades. By the 1950’s, people like Alan Turing, the British polymath involved in cracking the Enigma coding machine in WWII, explored the mathematical possibility of artificial intelligence. The field of AI was formally founded in 1956, at a conference at Dartmouth College, New Hampshire, where the term “artificial intelligence” was coined. For a more detailed read on the history of AI go here.

Since the 1950’s there have been early practical implementations of AI in the financial industries with applications to detect fraud.  In the ubiquitous online retail purchasing experience offered by retailers such as Amazon, AI has been used to suggest what else you should buy when buying an item. AI, like many technologies has had an evolutionary rather than revolutionary introduction into our daily lives. So why now? Why is it that now, AI is earmarked as an area of investment in so many companies across so many industries.

Today, companies have the right conditions allowing them to adopt AI solutions and have real business problems to solve. Companies have access to all the computing power, memory and storage they need for most complex AI scenarios.

So, it has become practical to deploy. And more importantly, companies produce copious amounts of data. Every company has a “big data” problem. Millions of data lines that touch on every aspect of their business, shipments, orders, point-of-sale, inventory reports, compensation and incentives, demand signals and so on. The data is vast and so diverse that purely representing the data through analytics and business intelligence platforms is not enough. The meaningful segmentation of information that would enable business insights and actions would require so many permutations of business intelligence that they are simply beyond human consumption and beyond any practical option to maintain.

Companies can be content knowing there is a bigger picture than just seeing pieces of the picture they do not see or do not understand. Curtailing their efforts to optimize revenue or, they can enlist the help of AI.

What is an Effective Approach to Implementing AI?

Some companies view AI as simply another platform they need to select and implement, much like they did in the past with BI tools. However, the story of BI offers some cautionary tales for the future adoption of AI.

The general idea was good, avoid having every department go build their own niche solution and their own special way of doing things. Create a single standard for the entire company that would service everyone. While this approach did solve certain problems, it helped centralize data helping create one “source of truth.” It simplified the IT landscape, made it easier to support users, etc. This approach, however, had a drawbacks. Because BI was now driven from a central platform, it required each business group and function in the company to engage and create their own “views” and “insights” into the data. It relied on individuals to maintain and develop new reports. In many cases 80% of the use was for executive forum reporting and actual use by individuals throughout the organization remained limited. Coupled with the fact that a lot of data, but not all the data, was centralized limited the value some companies got out of their business intelligence investments.

What’s worse is a variety of business applications that came bundled with useful business reports that end users and managers could use were never “turned on” because the decision was to use the main BI platform. The assumption was that all that is needed is the data be extracted, and intelligent reports built. Again, those reports were not maintained and not used and were not embedded into the contextual business process the end users where following daily. Value was limited.

When examining the opportunities for AI tools and adoption, it would be wise for companies to consider a mixed play. There are many valid scenarios for a platform investment to service cross company processes and even departmental initiatives. However, it is critical, especially in the context of revenue impacting processes, to allow AI solutions to be embedded within the process to drive insights and actions within the business flow. Otherwise, the adoption and effectiveness of AI solutions will be limited.


How can AI help companies optimize revenue?

Before we go into a few sample scenarios of practical use cases for AI and how they can help optimize revenue and business performance, let’s clarify a few terms sometimes used interchangeably. These different names are private cases of Artificial Intelligence and all fall under its umbrella definition.

  • Machine Learning – an iterative process through trials and testing that improves an algorithm to produce more accurate results. Think of Amazon telling you what else you should buy.
  • Predictive Analytics – combines statistics, data mining and predictive modeling that can also be refined with Machine Learning to make predictions about future outcomes.
  • Deep Learning – a broader aspect of Machine Learning, using artificial neural networks to identify patterns across multiple data sources and apply them to a bigger picture conclusion.

The most effective AI strategy will leverage these three aspects in combination to drive effective results as described in the following scenarios.

Pricing to Win

Pricing is an area that can benefit greatly from Artificial Intelligence. When setting up new price books companies are making a decision that will impact their win rates and market share. AI also benefits the negotiating a specific deal price concession that impact win rates and margins. Companies have a treasure trove of information in their transactional data including orders, shipments, channel POS, quotes and contracts. However, when it comes to deciding about a special price concession, most companies may have a process to review and might look at margin impact but that is typically the extent.

Model N has been providing companies with a deal intelligence solution that allowing contextual analysis – looking at a current deal in the context of similar deals, understanding the price curve and doing “what if” analysis on the fly to drive better price concession decisions.

Now, harnessing the power of AI, new insights and guidance are brought into the picture giving accurate predictions on the impact of price decisions and the chances to win. Using machine learning and an advanced neural network factoring the customer’s past performance, and a variety of other considerations, Model N’s Deal Intelligence application has demonstrated over 95% accuracy in predicting a win or loss outcome in specific deal scenarios on massive data sets.

This allows companies to strike a good balance of retaining high quality high margin revenue while improving their win rates. This same model is going to be expanded into price management functions.

How Much Volume Will a Customer Really Buy?

Another related problem is knowing what volume a customer will purchase. Customers often demand significant price concessions with the lure of “making it up in volume.” Companies find themselves giving discounts only to find that the volume purchased at the end of the day was lower. When making price concessions using deep learning, you can provide a contextual prediction of what volume the customer will purchase. This allows companies to do “what if” scenarios in real-time to see the impact of potential price changes on the chances to win and the volume of business that will be won. When initially introducing this challenge in Model N’s Deal Intelligence, the neural network produce results that were 65% accurate. However, in short order the deep learning process quickly elevated the accuracy of the volume predictions between 85% and 90% accurate. This is a true quantum leap for deal managers making on-the-fly decisions.

Which Incentives Actually Work?

Many companies in high tech offer a variety of market development funds (MDF) and rebate programs often paying millions and sometime hundreds of millions in return for some level of performance by their channels. The number of programs can run from as low as a few dozens to well over 1,000 different programs. These programs have different goals, different terms and conditions, different terms for calculating earn outs and payouts. It is challenging for companies to manage these processes effectively, while engaging their channels to ensure accurate payout. It is extremely difficult for many companies to understand the specific ROI per program. Which program was “free money” for the channel? Which program drove specific business outcomes? And as companies design and implement new programs before introducing them to their channels, before activating them, before even routing them for an internal approval – how do companies project the expected value of this incentive program? This too is a great place where AI can help provide these insights.

Through deep learning and predictive modeling, AI can help companies correlate channel performance using POS and Inventory data to programs and build on past performance to model the predicted efficiency of a new program as it is being designed. Providing real time feedback to the teams developing new programs such as: what’s the value and expected return and how may change as different elements in the program are changed.

There are many more scenarios for AI that can help companies optimize their revenue outcomes. These examples around pricing, deal negotiation and incentive program management and design are illustrative of the practical benefits AI can bring. If you would like to learn more about Model N’s AI efforts, click here.




Subscribe to our blog

Subscribe to our blog

Join us and get all the latest news. Select your communication preferences, so we can limit our communications to relevant topics.



Start typing and press Enter to search

Candlestick stock exchange backgroundTired businessman in complex network of lines