How to Apply Machine Learning to Business Problems

Daniel Faggella

Daniel Faggella is Head of Research at Emerj. Called upon by the United Nations, World Bank, INTERPOL, and leading enterprises, Daniel is a globally sought-after expert on the competitive strategy implications of AI for business and government leaders.

How to Apply Machine Learning to Business Problems 3
This Emerj Plus article has been made publicly available for a limited time.
To unlock our full library of AI use-cases and AI strategy best practices, visit Emerj Plus.

It’s easy to see the massive rise in popularity for venture investment, conferences, and business-related queries for “machine learning” since 2012 – but most technology executives often have trouble identifying where their business might actually apply machine learning (ML) to business problems.

With new AI buzzwords being created weekly, it can seem difficult to get ahold of what applications are viable, and which are hype, hyperbole or hoax. At Emerj, our market research focuses on cutting through the AI hype, and helping innovation and strategy leaders make a better business case for AI. This includes both our AI Opportunity Landscape research with enterprise clients, and our Emerj Plus best-practices guides for consultants and vendors.

In this article, we’ll break down categories of business problems that are commonly handled by ML, and we’ll also provide actionable advice to begin a ML initiative with the right approach and perspective (even it’s the first such project you’ve undertaken at your company).

Best of all, we’ll reference real business use cases, along with quotes and perspectives about “how to solve business problems with ML” from our network of AI researchers and executives. By the end of this article, you’ll have a good idea as to whether any of your present business challenges might be handled well with ML.

Note: At the bottom of this article, I’ve listed a basic glossary of ML terms in simple language.

What Kinds of Business Problems Can Machine Learning Handle

1 – Is the prediction you’re trying to make (or decision you’re trying to make) complex enough to warrant ML in the first place?

If it’s possible to structure a set of rules or “if-then scenarios” to handle your problem entirely, then there may be no need for ML at all. Also, if there is no precedent for any successful outcome applying machine learning to the specific problem to which you’re developing, it may not be the best foray into the ML world.

For illustrative purposes, it will be helpful to list a number of well-established business use-cases for machine learning so that you (the reader) can churn up your own application ideas:

  • Face detection: It’s incredibly difficult to write a set of “rules” to allow machines to detect faces (consider all the different skin colors, angles of view, hair / facial hair, etc), but an algorithm can be trained to detect faces, like those used at Facebook. Many tools for facial detection and recognition are open source. Below is a video of facial recognition using MATLAB:

 

 

    • Email spam filters – Some spam filtering can be done by rules (IE: by overtly blocking IP addresses known explicitly for spam), but much of the filtering is contextual based on the inbox content relevant for each specific user. Lots of email volume and lots of user’s marking “spam” (labeling the data) makes for a good supervised learning problem.
    • Product / music / movie recommendation – Each person’s preferences are different, and preferences change over time. Companies like Amazon, Netflix and Spotify use ratings and engagement from a huge volume of items (products, songs, etc) to predict what any given user might want to buy, watch, or listen to next.
    • Speech recognition – There is no single combination of sounds to specifically signal human speech, and individual pronunciations differ widely – machine learning can identify patterns of speech and help to convert speech to text. Nuance Communications (maker of Dragon Dictation) is among the better known speech recognition companies today.
    • Real-time bidding (online advertising) – Facebook and Google could never write specific “rules” to determine which ads a given type of user is most likely to click on. Machine learnings helps to identify patterns in user behavior and determine which individual advertisements are most likely to be relevant to which individual user.
    • Credit card purchase fraud detection – Like email spam filters, only a small portion of fraud detection can be done using concrete rules. New fraud methods are constantly being used, and systems must adapt to detect these patterns in real time, coaxing out the common signals associated with fraud.

(For lists of specific industry applications, explore our other articles about ML in marketing, ML in healthcare, ML in robotics, and ML in finance)

2 – Do you have new data and clean data?

“Clean data is better than big data” is a common phrase among experienced data science professionals. If you have reams of business data from years ago, it may have no relevance today, particularly in fields where the basic business processes change drastically year-over-year, such as mobile eCommerce). If you have reams of unstructured and disjointed data, you may have too much “cleaning” to do before you can ever get around to learning from the information collected.

UBER’s Head of Machine Learning Danny Lange once recommended that companies just starting out in machine learning should begin by applying supervised machine learning to historical data. Find data that’s already clean and relatively recent, and use labelled training data to start finding insights.

Note that in a rapidly-changing field, newer data is positively required. For example, if you run a door delivery service for pet supplies, and your app, prices, product offerings, and service areas have changed significantly over the last six months, you will need much more recent data to learn from than, say, a company selling homeowners’ insurance in Montana. If data is not related to the relevant trends and nuances of your current business, it is unlikely to glean predictive value.

3 – Does your data have existing labels to help a machine make sense of it?

While unsupervised learning (see glossary below) allows for a wide degree of applications in making sense of data without labels, it’s usually not advised for companies to “jump into” ML with a first application in unsupervised learning. The low-hanging fruit for an ML use case is likely to spawn from its historical, labelled data. Below are some examples that might help a reader garner new ideas:

  • Facebook had millions and millions of tagged human faces on its platform, faces that were already associated with an individual person. This gave Facebook the ability to train algorithms on a tremendous volume of labeled data, with millions of faces in all kinds of lighting conditions and from various angles, allowing the algorithms to be highly refined and attuned to identifying specific human faces.
  • Google serves billions and billions on search results, and can gauge the usefulness and relevance of its search results based on click-through rate of its top lists, page -load time, time-on-page from a specific visitor, and many other factors. It would be impossible to find a set of hard and fast rules for showing the right search results, so Google’s algorithms learn what the best options will be based on real-time engagement from billions of daily searches.
  • Credit card companies like CapitalOne are faced with a huge volume of chargebacks and reportedly fraudulent purchases each day. By finding connections and patterns across types of purchases, locations of purchases, and types of customers, CapitalOne can use the “labelled” instances of fraud to predict other transactions that are most likely to be fraud. Anomaly detection plays an important role in various security and fraud applications; below is a short “anomaly detection primer” video from Numenta’s YouTube channel:

 

 

      • An eCommerce company with a massive volume of customer support emails will have a lengthy record of support tickets that were labelled “refund requests”, “technical issues”, “delivery issues”, among others. The company may choose to develop an ML system to instantly label incoming emails, transcribed phone calls, and chat requests with the proper support issue “type.”

(See the comments from the Emerj network for more ideas about using labelled data)

4 – Can your solution to this problem afford for some allowance of error?

ML might be thought of as a kind of “skill”, in the same sense that one might apply the word to human beings. A skill that’s alive, adapting, growing and informed by experience. For this reason, an ML solution will often be incorrect a certain percentage of the time, especially when it’s informed by new or varied stimuli. If your task absolutely cannot allow for any error, ML is likely to be the wrong tool for the job.

An example of an application that cannot allow for error might be an application that aims to read the amount of an invoice or bill and then pay that invoice or bill. One letter difference or one number difference could mean overpaying your bill by 10x the original amount (if the decimal was interpreted to be in the wrong place), or sending money to the wrong company (if an invoicing company name isn’t registered exactly).

In a case like above, some degree of ML might help with “bucketing” different types of bills or invoices, but the final decision to enter the payment amount and send a payment would likely require an accountable human.

As an interesting caveat, there is a San Francisco-based startup called Roger.ai which is aiming to use natural language processing and machine vision to real and pay bills, albeit it pulls humans into the loop before sending funds.

Quotes from the Emerj Network:

In order to gain additional perspective on the issue of “picking a business problem for machine learning”, we decided to reach out to our network of previous AI podcast interview guests for additional guidance for our business readers:

Dr. Ben Waber — PhD, MIT — CEO of Humanyze (AI-powered people analytics company):

Any business problem where you have hard data, variability, and a large number of examples.”

Dr. Danko Nikolic — PhD, University of Oklahoma — Data Science and BD&A, Computer Sciences Corporation:

“1. There is a lot of savings to be made by companies.

2. Predictons require many variables, complicated non linear relationships between them and in some cases are highly stochastic. So it is often only algorithms that can learn those relationships. Humans alone would have hard time.”

Dr. Charles Martin — PhD, University of Chicago — CEO, Calculation Consulting:

The best problems are those in which there is a very large, historical data set that includes both rich features and some kind of direct feedback that can be used to build and algorithm that can be implemented and tested easily and will either decrease operational costs and /or increase revenue immediately.

Dr. Ronen Meiri — PhD, Tel Aviv University — CTO & Founder, DMWay

“Most business problems can be appropriately addressed using two machine learning methods:
1st: ‘What will likely happen?’ and
2nd: ‘What is the future expected value of …?’.

The first is a classification type problem that includes classifying who is likely to Churn, Default, Buy, Sell among many others use-cases.

The second question is an Expected value problem that is solved by regression and gives accurate predictions for a variety of use cases like Pricing Optimization and predicting Life Time Value.”

Peter Voss — CEO, AGI Innovations Inc:

(To begin, Peter quotes Dr. Robin Hanson, Professor at George Mason University: “Good CS expert says: Most firms that think they want advanced AI/ML really just need linear regression on cleaned-up data.”)

“I think that most businesses don’t justify the investment in ML/DL (of course, ML means many things). Cutting edge stuff that everyone is talking about requires a lot of data and expertise, and is static – i.e. it needs to be retrained when data or categories change.”

Linear regression for business problems

Linear regression is one of the oldest, simplest, and widely used machine learning models. Some researchers contend that many intermediate prediction problems may need little more than this basic approach, at least initially. Image courtesy of MathWorks.

Peter’s comment is apt, and shouldn’t be taken lightly. When ML is sought out because it’s hip and popular, it’s unlikely to yield significant results. Find the tool that best suits the needs of your bottom line; there’s a high likelihood that ML may not be the solution you need to meet your business or growth goals.

This leads us into the second major section of this guide:

Pointers for Applying Machine Learning to Business Problems

1 – Begin with a priority problem, not a toy problem

In an off-mic conversation with Dr. Charles Martin (AI consultant in the Bay Area), he mentioned that many companies read about ML with enthusiasm and decide to “find some way to use it.” This leads to teams without the real motivation or gusto (or committed resources) to drive an actual result. Pick a business problem that matters immensely, and seems to have a high likelihood of being solved

UBER’s Danny Lange has mentioned from stage that there is one thought process that’s highly likely to yield fruitful machine learning use case ideas: “If we only knew ____.”

Ask yourself, what mission-critical business information are you dying to know, but can’t currently access? Maybe it’s understanding the lead sources most likely to yield the highest customer lifetime value, or the user behavior most indicative of expected churn.

Emerj Professional Membership 1200x200 Copy@2x

2- You can give it data, but all of the context must come from you

Thinking through what information to “feed” your algorithm is not as easy as one might presume. While ML algorithms are adept in identifying correlations, they won’t understand the facts surrounding the data that might make it relevant or irrelevant. Here are some examples of how “context” could get in the way of developing an effective ML solution:

  • Predicting eCommerce customer lifetime value: An algorithm could be given data about historical customer lifetime value, without taking into account that many of the customers with the highest lifetime value were contacted via a phone outreach program that ran for over two years but failed to break even, despite generating new sales. If such a telephone follow-up program will not be part of future eCommerce sales growth, then those sales shouldn’t have been fed to the machine.
  • Determining medical recovery time: Data might be provided to a machine in order to determine treatment for people with first- or second-degree burns. The machine may predict that many second-degree burn victims will need only as much time as first-degree burn victims, because it doesn’t take into account the faster and more intensive care that second-degree burn victims received historically. The context was not in the data itself, so the machine simply assumes that second degree burns heal just as fast as first degree.
  • Recommending related products: A recommendation engine for an eCommerce retailer over-recommends a specific product. Researchers only discover later that this product was promoted heavily over a year ago, so historical data showed a large uptick in sales from existing buyers; however, these promotional purchases were sold more based on the “deal” and the low price, and less so by the actual related intent of the customer.

3 – Expect to tinker, tweak, and adjust to find ROI

Building a ML solution requires careful thinking and testing in selecting algorithms, selecting data, cleaning data, and testing in a live environment. There are no “out-of-the-box” machine learning solutions for unique and complex business use cases. Even for extremely common use cases (recommendation engines, predicting customer churn), each application will vary widely and require iteration and adjustment. If a company goes into an ML project without resources committed to an extended period of tinkering, it may never achieve a useful result.

Quotes from the Emerj Network:

We again reached out to our network of Emerj interviewees and consensus respondents for opinions and tips on implementing machine learning in business. Below are a collection of quotes:

Dr. Ben Waber — PhD, MIT — CEO of Humanyze (AI-powered people analytics company):

You cannot use ML to solve business problems in a vacuum. Make sure you get buy in from business unit leaders to make concrete changes based on the analysis.”

Dr. Danko Nikolic — PhD, University of Oklahoma — Data Science and BD&A, Computer Sciences Corporation:

The most common mistake that businesses make when using ML is that they think that an ML solution is a one-shot process: They send data to data scientists, and data scientists send back THE model. In contrast to that, finding good ML solution is an iterative process that involves research, trials and errors, experimenting, talking to the business experts, etc.

ML cannot ever become commodity. Success of ML depends strongly on the knowledge, skills and dedication of the people who do it.”

Dr. Charles Martin — PhD, University of Chicago — CEO, Calculation Consulting:

“Avoid setting up massive infrastructure until you have a handle on what you want to do.  You can easily spend 6 months to a year setting up Hadoop and Spark and not see any ROI.

You will be lucky if 5% of your data is correct and useful.  You need to design an experiment that can identify the low hanging fruit and ferret out the data you need.   You can build an algorithm on a high memory AWS node.

Get the algorithm into a live environment and test it as early as you can. Don’t build try full production system.   Remember, ML is about math, not coding !  You want to test it small.  Run enough examples to flush out the problems, but not so small that the statistics are meaningless.”

Ferris Jumah — Previously ML at LinkedIn — Bay Area ML Consultant:

“Get data driven as soon as possible. Machine Learning doesn’t come for free. You need to build intuition around your data, how you measure the business and know your customers, link not just measurements but also insights to decision making. Log everything, build storage and processing systems, ensure they are accessible, conduct deep analysis and as many experiments as you can on your product, build in intelligence into as much as your product as possible.

At this point, consumers expect personalization and “smart” features. Build them in, learn from them, and ensure that you have a feedback mechanism in place. Finally, hire and invest in data people who are passionate about your problem and business.”

The consensus (in the limited number of quotes above, and from dozens of other conversations with business-minded data scientists) is that machine learning is not as much of a mere “tool” as, say, marketing automation software. Anyone with a good manager and a bachelors degree from a community college can pick up Constant Contact or even (with a bit of tinkering and calling the support line) Marketo or Hubspot and drive some company value.

machine learning workflow

There are no simple shortcuts to iterative, multi-faceted process of applying machine learning. image credit Microsoft’s CortanaIntelligence.

 

ML doesn’t yet show up in a neat box, and value is wrought by hard thinking, experimental design and – in some cases – hard mathematics. A little bit of time on Google and YouTube, and you can get a hang of how to set up DropBox for your business. Predicting churn rate across your customer segments with machine learning? Not the same game.

Preparing to derive business value from ML implies having trained talent, expert guidance, and an (often) tremendous “data cleansing” period – and none of it is guaranteed to be a win, as Dr. Martin states aptly above. If Google, Amazon, and Facebook could get their interns to set up ML systems, would they really be spending millions and millions of dollars to scoop the world’s top AI talent out of academics to work for them?

While machine learning isn’t an easy setup, it’s also not one that any future-minded business can leave off the table for too long. The efficiencies gained by the “rockstar” tech companies through machine learning are substantial, and startups here in the Bay Area aren’t just getting funded because “machine learning” is a buzzword – it’s also because many of them have a powerful and strong business case.

Interested readers might benefit from our recent consensus of 26 machine learning / AI researchers where we asked: “Where should machine learning be applied first in business?” The infographic featured drives home many of the same points highlighted in this article.

The ultimate question for executives remains: When can we have (a) the resources required to invest in machine learning seriously, and (b) a legitimate use case that started from trying to find real business value, not from “trying to find a way to kinda use machine learning.” That’s a thought process that can’t be done for you, but our hope is this article has helped to inform your perspective and give you resources to draw from in future.

Glossory of Machine Learning Terms

Unless otherwise noted, definitions have been sourced from Wikipedia.org.

  • Machine Learning (ML) – Machine Learning is the science of getting computers to learn and act like humans do, and improve their learning over time in autonomous fashion, by feeding them data and information in the form of observations and real-world interactions. (definition taken from our “What is Machine Learning?” guide)
  • Supervised Learning – The machine learning task of inferring a function from labeled training data.
  • Unsupervised Learning – The machine learning task of inferring a function to describe hidden structure from unlabeled data.
  • Classification – Identifying to which of a set of categories a new observation belongs, on the basis of a training set of data
  • Regression – A statistical technique for estimating the relationships among variables (includes linear regression, logistic regression, and other approaches)
  • Algorithm – A self-contained step-by-step set of operations to be performed. Algorithms perform calculation, data processing, and/or automated reasoning tasks.
  • Natural Language Processing (NLP) – A field of computer science, artificial intelligence, and computational linguistics concerned with the interactions between computers and human (natural) languages

A Thanks to Our Machine Learning Respondents

I’d like to extend a special thanks to our respondents for this extended article. Below you’ll see links to hear our full interviews with these researchers and businesspeople, as well as links to their respective organizations:

 

Image credit: ExtraordinaryCEO