Spark Sales

with the best recommendation engine
Boosting Conversion Rate with


The Recommendation Engine on a website is the embodyment of the best salesperson, only scalable. The best practices implemented to deliver superior recommendations are:


Understand Customer Individual Needs (Tradeoffs and Preferences)

The best way to gain customer satisfaction is to use a technology that mimics the way they think when they buy. Every decision we make, we consider tradeoffs. It is universally accepted as the natural decision making process. We use the best data we believe that we have, tempered with our preferences and weighed by our tradeoffs to select the optimal choice or solution.

Engage and Interact with Customers

The web is by nature INTERACTIVE. AI/ML engineers have often shown a tendency to wan to guess in a single attempt the best product for a customer. There is great value in engaging the customer and asking them about their purchase driving factors.

Find the Best Solution for Each Customer

This process is highly personalized. It is based on the specific expressed preferences and the importance a customer associates with various capabilities of the solution
"He who establishes his argument by noise and command shows that his reason is weak" Montaigne




There are several technologies needed to mimic the brain: 

  • At the core is the TRADEOFF engine that works the way our brain makes decisions. It weighs the various criteria that are taken into account to reach a decision.
  • Since it is not always easy to explicitly articulate one's tradeoffs, our COMPARATIVE engine infers the tradeoffs based on some sample ranking. The comparative engine can take a set of ranked products and measure what tradeoffs would yield the given ranking. The combination of the Comparative Technology with the Tradeoff engine yields the best recommendation every time. The comparative technology also allows a customer to compare side by side two or more products and highlight their differences.
  • Our mind occasionally uses HIERARCHICAL and parametric searches when the target solution is known. i.e. when searching for "Nike Air Force 1" the easiest way to find it is to use a hierarchical appproach.
  • While PARAMETRIC searches are suited for situations where one or more constraints are absolutely required.
  • ASSOCIATIVE technology uses social graph theory to discover new relations we often call serendipitous, but in reality they have a scientific explanation.
  • Finally, our brain is phenomenal at RECOGNIZING PATTERNS and AI/ML is at its infancy stage in being able to do so. Current AI/ML requires large volume of data that our brain cannot handle. The types of volume that few organizations collect (Google, Amazon, Microsoft, Apple).
  • AI/ML 2.0 addresses the DATA VOLUME APPETITE of traditional AI/ML systems and makes it more humanlike in that sense that it can apply intelligence to a more reasonable data set of products.
  • One question that often surfaces is tied to the ability to tackle the PSYCHOLOGICAL aspects of buying, or EMOTIONAL INTELLIGENCE. A Tradeoff based recomendation model is designed to take into account psychological factors too, by enabling the incorporation of these factors as criteria to tradeoff. For example "Keeping up with the Joneses" is always a consideration and is weighed in relation to the other characteristics. In summary, the advances in technologies that mimic the way the brain works are at the core of the ability to recommend better and sell more.



How to assess a Recommendation Engine


What criteria are used to assess Recommendation Engines and how do they correlate with the conversion rate?

Results Quality
The primary criteria that impacts conversion rate is the results quality. Taking into account the factors that matter to each customer and their respective importance yields the best results.
Results Quantity
Avoiding confusion with too many choices or options is the second most important factor in maximizing conversion rates. Presenting the right number of choices is essential.
Gaining insight on the customer's preferences and tradeoffs (i) enhances the custmer journey and experience, (ii) boost the cross-sale and Up-sale .
"If you always do what you always did, you will always get what you always got" Einstein

Comparing Tradeoff Recommendations with Older Techniques

"If not us, who?" JFK

About Us

Sparkdit is a division of VITEDS, LLC. Our mission is to build the uncontested best recommendation engine.

  • Leverage the Humanlike Intelligence Platform
  • Accelerate Deployment of Solution
  • Yield Return on Investment by Exceeding Target Conversion Rate
"By working together, pooling our resources and building on our strengths, we can accomplish great things" R. Reagan

Our Team

Our business is about innovation and our greatest asset is our People.


Fadi Micaelian - CEO

MIT, INSEAD - Oracle, BroadVision, Auguri, Intellectual Ventures


Ed Zanelli - SVP Engineering & Architect

Caltech - Oracle, Sybase, Gain, Siebel, Citrix


Stefano Gargioli - VP Sales

La Sapienza - IBM, Cisco, SAP, BroadVision


Sanjay Prasad - General Counsel

BU & Syracuse - Fenwick & West, Oracle, IPValue, Intellectual Ventures, Prasad IP


Peter Chu - Chief Strategist

Stanford, Harvard - Oracle, Verity, ApaceWave, BroadVision


Charles Yeterian - CFO

Univ. of Dallas - Booz Allen, BNY Mellon, Novus Aviation, Aviation Sans Frontières


Dr. Pehong Chen - Advisor

E-Commerce Pioneer, Founder and CEO of Gain Technnology and BroadVision


Court Lorenzini - Advisor

Founder and CEO of DocuSign


Ed Oates - Advisor

co-Founder Oracle

"Those who dare to fail miserably can achieve greatly" JFK


"Every accomplishment starts with the decision to try" JFK

Contact Us