top of page

Using Forward-looking Algorithms to Drive Businesses Forward

Algorithms have been part of the worlds most successful companies such as Google. With organisations now generating more data than ever, collecting data is no longer a problem – instead, it’s all about how that data can be used and processing data is done best when it’s achieved through algorithm.

An algorithm is defined as a procedure or formula for solving a problem, based on conducting a sequence of finite operations or specified actions. It extracts systematic patterns from data combined with business knowledge to predict future events or behaviours. It uses simulated scenarios to discuss possible options and decide on any necessary action.

Generally, it reduces the number of tedious manual tasks and enables people to focus on value-add activities. At the same time, it generates smarter insights and stronger business outcomes through simulations and optimisation.


How does a forward-looking algorithm work?

Most companies traditionally use a backward-looking algorithm. Forecast solely relies on past data (no real-time data). Inputs are manually programmed based on past data.

A forward-looking algorithm however enriches data from the backward looking model. It gathers huge amounts of data in real-time and merge it with existing data. Analytics experts cast complex judgement on future business prospects by applying deep business understanding.

For example, each user is shown recommended titles at Netflix. The dynamic recommendation system puts together data collected from various places. For example, other members with similar taste and preferences. Every time a user chooses a movie, the algorithm is updated. The algorithm then generates recommendations for users.

Netflix Recommendation, algorithm, recommendation algorithm
Netflix Uses Algorithm for Its Recommendation System (Source: Proof Blog)

Other examples include optimising delivery routes and creating dynamic pricing for airline carriers to adjust the ticket price as per the demand-supply ratio to maximise profits. Real-time data such as weather, traffic conditions, customer characteristics (past booking) are used in the algorithm.

It communicates business intelligence in real time. With a stable and mature algorithm, decision making can be automated.


How to build a reliable algorithm?

A good understanding of the management's question is required to define the right questions the algorithm should create the output on. For example, a large IT hardware company with large portfolio of items, may ask the question "If I pick 100 products, which ones should I choose to maximise revenue from orders containing only these products?". Questions have to be specific for the output to be actionable.

Managerial questions will then be converted into solvable statistical questions. This process is not easy, as it requires good statistical knowledge and business acumen. In the example, the scenario maybe complex as there can be different combinations of IT hardware such as power supply, server, computers that a customer will order.

The process also depends critically on knowledge of strategic new-product introductions and planned obsolescence, which historical order data does not reveal. Next step is to determine if data is available and sufficient to answer the questions. The quality of insights generated depends on the quality of source data. It has also become easier to acquire data.

business knowledge, business knowledge and data, algorithm, data algorithm
Combining business knowledge and data is critical to build the algorithm (Picture credit: Pexel)

The right algorithm with good reliable results/ output can then be built for the business setting. Otherwise, questions need to be re-framed and the entire process repeated. Having the right questions for the forward looking algorithm define the overall success of a business. Data is only as good as the questions asked.

Based on the question, available data, and experts inputs, the algorithm then estimates the probability of outcomes and guides decision making.

To get there, companies must ensure that employees develop trust in the accuracy and stability of the algorithm and related processes over time, so the model will be used in their day-to-day work.


How can the algorithm remain relevant?

Every algorithm project should start small. With every small win, more questions should be asked and more data added for use. More features can be derived from existing variables, where hidden relationship in a data set can be unleashed.

The performance of the algorithm should always be improved through additional statistical analysis. Algorithms can then be further developed, refined and tuned. There is a need to continuously improve the outcomes of those patterns through “learning.”


Those companies that are willing to use algorithms to provide insights will see its efforts translate to the bottom line. When primary factors that drive a company's success are measured, closely monitored and predicted, the company is in a much better position to adjust, advance and mitigate risks. This is especially important in a fast-moving world.

Do you use algorithms in your business? If so, how do you use it? Share by leaving us a comment. If you require more information or assistance to develop algorithms, contact us. We want to be an extension of our clients. Subscribe to our newsletter for regular feeds.

Did you find this blog post helpful? Share the post! Have feedback or other ideas? We'd love to hear from you.



Arrfelt M. et al (2013), Looking Backward Instead of Forward: Aspiration-Driven Influences on the Efficiency of the Capital Allocation Process, The Academy of Management Journal, Vol 56 No 4, pp. 1081 - 1103

UX Planet, Netflix: Binging on the Algorithm,, published 2 August 2018

141 views0 comments


bottom of page