AI, IoT, M2M, Big Data – The Alphabet Soup of Technology Jargon You Need to Understand (Part 1 of 2)

In 1984 I was seventeen years old and working as an usher in a movie theater when the science fiction thriller The Terminator was released. It was a surprise hit, and I must have seen the movie a couple of dozen times. In case you are not familiar with the movie, Arnold Schwarzenegger plays a human-like cyborg, a Terminator, sent from the future with a mission to kill Sarah Connor, the mother of the future resistance leader that is fighting the Terminator’s artificial intelligence master, Skynet. Aside from the obvious standout qualities of Schwarzenegger’s physique (a former Mr. Universe and Mr. Olympia) and the incredible strength demonstrated by the cyborg, the Terminator looks and even acts somewhat human. To remind the audience that the Terminator is actually a very sophisticated computer, director James Cameron sometimes displays the action from the perspective of the Terminator.

In these “look through” scenes, the audience is presented with a screen that is apparently the field of vision of the Terminator. The film color quality is replaced with mostly red, white and black imagery. Superimposed on the imagery is a bunch of scrolling text gibberish and some highlighted, flashing square boxes to call attention to certain data elements the Terminator may be analyzing – a person’s body size for suitable clothing, weapons in the hands of potential antagonists that must be foiled, etc. Of course, if the Terminator was really a sophisticated computer cyborg, there would not be an internal display barfing computer gibberish onto a screen in a manner that was readable by humans. Computers do not need human-readable text to operate on data the way humans need it. The computer would simply be ingesting external data via the cyborg’s camera eyes and his microphone ears along with any other external sensors for temperature, pressure, odor, and what not. Based on this observed data, the Terminator would be making judgments and taking actions that would have a high probability of creating a path to accomplish the mission – the termination of Sarah Connor. All of this would be happening without a human readable screen display.

Why am I talking about The Terminator? Why is the detail of the Terminator’s view of the world as depicted by the movie director important? I am talking about The Terminator to illustrate the point that artificial intelligence (AI), the Internet of things (IoT), big data, and all of the other alphabet soup puked up on a daily basis by technology media and vendors hyping their products is generally nothing more than the collective, gradual evolution of computers. In 1984, James Cameron could imagine a computer that understands and speaks natural language, sees real-time imagery, reacts to its environment, and takes actions to accomplish the mission. To portray the Terminator as a sophisticated AI being, Cameron showed the audience a visual model that generally represented what computers looked like to the masses in 1984 – a somewhat low-resolution screen with digitized text scrolling on it with an occasional selection option that would become highlighted if you tabbed a cursor to it (remember, the mouse was a new thing in 1984 as the first Apple MacIntosh computers just shipped that year). Cameron could not assume that the audience would make the leap to his futuristic interpretation of an AI-enabled cyborg, so he showed the audience a 1984 computer interface to make certain they got the connection. All this stuff in the media about AI, IoT, machine learning, big data, blah, blah, blah is just the real world catching up to what James Cameron predicted would happen way back in 1984.

Today we are talking to our phone to have it dial our best friend. We are issuing verbal commands to our Alexa assistant to have it order pizza or play our favorite music. Our Nest thermostat is monitoring our habits, such as when we come and go, along with our preferences for ambient temperature in order to take actions regarding raising and lowering the temperature where we live. These common applications of AI would have been totally foreign and inconceivable to a movie audience in 1984. But James Cameron had a vision of what artificial intelligence could potentially accomplish in the future, and he did a really good job presenting that vision to the audience in a way that they could understand it. Let’s do a quick reset on some over-hyped terms – AI, IoT, and big data.

Artificial Intelligence – AI is just the trend toward computers ingesting more diverse data in more formats (i.e. images, audio, natural language, pressure, temperature, humidity, etc.) to enable analysis that leads to judgments and actions related to accomplishing a mission or objective. Because AI is more of a trend than a definitive end-state, AI can simply be classified as Hofstadter, a famous AI scientist, describes it – “AI is whatever hasn’t been done yet.” More accurately, AI is simply the leading edge of new capability for computers to operate more intelligently on a broader diversity of data.

Internet of Things – IoT is simply a trend where more and more things are connected to the Internet to send or receive data or to act upon data received. Historically connections to the Internet were people staring at screens (and increasingly listening to audio speakers) and entering data or responding to data received. “Things,” whether a cyborg like the Terminator or a $10 temperature sensor, don’t need screens (nor keyboards or a mouse or speakers) to send and receive data or to act upon data received.

Big Data – Big Data is simply the collection and analysis of data sets that are too large for humans to effectively parse, analyze, and extract intelligence from using simple programs like Excel. Ever cheaper storage and computing cycles lead to ever-increasing data collection, storage, and analysis. Again, big data is simply a trend and not a definitive end state.

Over time, computers will progress to read a broader spectrum of inputs, make more sophisticated judgments, and take an increasing variety of actions that lead to desired outcomes. No one was talking about AI in 1984 – no one in the mainstream media anyway – because the topic was confined to a small group of computer nerds at top technical institutions like Stanford and MIT. Yet the director of The Terminator could imagine a future where a computer becomes so powerful that it can measure its environment in a humanlike manner, make judgments based upon those measurements, and take intelligent actions to execute a mission – in this case, the termination of Sarah Connor. It is unlikely that anyone who saw The Terminator in 1984 remembers the on-screen effects that Cameron used to connect the audience to the idea that the Terminator was a computer. I bet everyone who saw the movie remembers the Terminator’s mission, however. What was the mission? To terminate Sarah Connor of course.

Whether an innovation can correctly be labeled as AI (or with any other overhyped term of the day) is far less important than whether the innovation helps accomplish the mission. The Terminator’s mission was to terminate Sarah Connor, and the Terminator was extremely well suited for carrying out the mission (although it actually failed in this case). Defining the mission that you would like to accomplish with AI, IoT, big data, etc. is actually much more important, in my humble opinion, than the actual technology you select to achieve the mission. Have you thought about the mission that you want to accomplish using technology?

I believe the mission you are generally attempting to accomplish through technology is to maximize customer equipment performance while eliminating equipment failures so that your customer experiences the least risk, expense, and disruption in their business. The reason that technology is important as an enabler of this mission is because it is generally cheaper and easier to manage (sometimes) than people. If you accomplish this mission, your customer will spend zero dollars recovering from disruptions (lost output, spoiled inventory, damaged property, emergency services) while maximizing the amount of money they spend with you relative to other suppliers.

I will continue this topic next week with real-world examples of how AI, IoT, and big data are being used by service contractors today. I’ll also have some advice to make sure that what you’re building is more like a high-performing Terminator than a cobbled-together Frankenstein monster.

Continue to part 2 here.

What’s Technology Worth? How to Value Your Technology Investments.

I hear customer prospects cry out for “the perfect application for my business that does everything” in nearly every sales call that I make. It does not exist. I have argued again, and again, and again that every business of any size will ultimately buy multiple applications to serve the diverse needs of their business functions.

Look at your phone. One application? Or many? Displaying the weather is different from transferring money from your bank account is different from measuring the intensity of your workout is different from keeping up with your social network.

Likewise, your accounting function is different from your sales function is different from your customer service function is different from your marketing function. The idea that one application will be sufficiently good for your business to remain competitive in all of these different functions is silly, and any software vendor promising you that outcome is a silly vendor. But what about the follow-on questions:

• If I am going to buy many applications, how much should I expect to spend?
• How do I value applications that make my business more competitive in a world where technology innovation increasingly determines market competitiveness?

Guideline 1: Select Modern Software With Open APIs

Well, before you even consider how much to pay, you need to perform the first and most basic test in the software buying cycle. Go to your favorite online search engine and enter the following query:

[INSERT NAME OF SOFTWARE APPLICATION HERE] API documentation

The first organic link below all of the advertisements from the software vendors that are trying to sell you a competing application should be a link maintained by the vendor of the application in question. That link should lead you to detailed documentation for how the application you are considering can be integrated with other applications that you use. Application Programming Interfaces (APIs) are the key to a new world of connected innovations for your business. Without good APIs that are publicly documented, the application you are considering is worthless. You should not pay anything for it.

Go ahead and try the search for a couple of high-quality applications that are on the market today. Insert “ServiceTrade” or “ZenDesk” or “PipeDrive” or “Marketo” or “Hubspot” or “Slack” into the query above. Check out the first organic link below the advertisements. What do you see? This query is the first test to determine if an application is worth at least a penny.

Guideline 2: Judge the Technology’s Value to Your Business

Let’s say that your application passes that first test. What now? How much is it worth? Well, it sort of depends on how much it increases the value of your business. In a prior blog post, I argued that the questions that determine the value of your business are How Many? How Much? and How Long? How many customers do you have and how many can you attract with your value proposition? How much can you charge those customers for the services that you provide to them? How long can you keep those customers when you are charging a significant premium compared to your low price competition? These are the questions that you should use to evaluate how much a new software application is worth to your business. The more the software impacts these measurements, the more you should be willing to pay because it is going to make your business more valuable.

Does the new application help me attract new customers? Does it help me charge them more because it provides my service with some new features that customers value? Does it help my business become sticky so that it is difficult for customers to fire me and replace my service with a low-cost competitor? If the answer to these questions is “yes, absolutely, definitely” then the application is probably very valuable. If the answer is “no, not really” then the application is only worth some fraction of the money it might help you save by eliminating administrative burden.

The problem is that you are probably significantly overpaying for administrative applications like accounting and underinvesting in applications that drive new customer acquisition, service differentiation, and revenue. And I also bet your accounting application provider is telling you “we have a plugin for sales, and customer service, and technician management, and every other thing you might need” in order to justify the crazy price you are paying for that application. Am I right? Probably.

So how can you alter your portfolio of applications through time to push down the expense associated with administrative applications so that you can reinvest those dollars in applications that actually drive up the value of your business to its shareholders? Applications that enhance your ability to add customers, charge them more for your services, and hold onto them longer?

Conclusion: Recommendations

My first recommendation is to only consider modern software as a service (SaaS) applications that have publicly documented APIs. These will generally be cheaper than the older, legacy server-based applications, and they will deliver more innovations to your business going forward. Software investors are NOT investing any of their precious capital in old server applications, so these legacy applications are going to stagnate and die. No point in throwing your money away on a dead horse.

My second recommendation is to ask the basic questions around How Many? How Much? and How Long? for new applications you are considering. If the applications you are considering do not contribute to these value metrics, then simply look for the low price alternatives that meet the SaaS and API criteria and determine how much administrative expense they might save you. You can spend up to 100% of the savings on the administrative applications to eliminate manpower spending.

If the applications do in fact help you attract more customers, charge them more for valuable new features, and hold onto them forever, open up the wallet and let fly for up to 2 – 3% of the revenue you expect to drive by being the most innovative service contractor in your market. I assure you that the best service contractors will collect a 15 – 25% revenue premium in their market, which easily justifies the spending on the applications that drive that differentiation. I will also assure you that competing on technology innovation is much more fun than competing on price.

Case Study: Investing 3% of Revenue in Technology

Let’s look at some examples from ServiceTrade’s business to set some benchmarks for how much to pay.

The biggest technology application expense category that ServiceTrade faces is for infrastructure services that power our customer’s experience with our product. Amazon and Google charge us for technology that provides neat features in our application. The ability to send a quote to a customer via an email with a link that presents the quote online with photos and video and audio and a “one click to approve” button that drives revenue for our customers is largely dependent upon capability provided to ServiceTrade by Amazon. The ability to map customer locations for scheduling efficiency, see the locations of the technicians in real time, and prefill the fields for setting up new customer location records is largely dependent upon capability from Google. The applications from Amazon and Google are VERY valuable to ServiceTrade because they help us attract new customers and charge them a premium, and we spend about 6% of our revenue on these types of applications.

Now, ServiceTrade makes about 80% gross margin on the applications we sell, so we can afford to spend heavily on making these applications great. If your service to your customer drives a lower margin, say 35%, then 6% of revenue makes no sense for any technology. The apples-to-apples comparison, in this case, is probably close to 7% of gross margin (roughly), which would equal 2.6% of revenue for an application that really helps you deliver differentiated value to your customer. So for a $10 million dollar service contracting business generating 35% gross margin, the equivalent amount would be $260,000 per year.

The next biggest category of technology expense at ServiceTrade is for sales and marketing applications. We have Salesforce, Marketo, Salesloft, and a handful of other applications that help us present our value proposition to customers in a way that drives new sales. These applications help us increase the How Many customers metric. We spend about 1.5% of revenue on these types of applications. Again, to adjust for gross margin, that would be about .6% of revenue for a 35% gross margin business. So for a $10 million dollar service contracting business with 35% gross margin, the equivalent annual expense would be $60,000.

The next biggest category of technology expense at ServiceTrade is for customer service oriented applications. These are the applications that help our engineers and our support staff keep track of how things are going for our customers and to monitor the application for errors or potential signs of trouble. We spend about .4% of revenue on these types of applications. They are tangentially oriented toward helping with the How Long can we keep our customers question. Clearly, these are far less valuable than Google and Amazon, and also less valuable than the sales and marketing applications, both of which help us drive up the How Many? and How Much? elements of our business value. Adjusting for gross margin again, and you get .16 as the percentage of the revenue in a 35% gross margin business. A $10 million service contracting business should consider spending $16,000 per year on customer service infrastructure.

Finally, there are the administrative applications like accounting, email, file sharing, calendar, reporting, office productivity, etc. These are the applications that every business needs, but their value is simply in keeping the administrative burden of running a “tight ship” as low as possible. ServiceTrade spends about .3% of revenue on these type of applications, and it is unlikely that the expense of these will scale linearly as we grow. When we double in size, I would expect that percentage of revenue to be about .2%. So for a $10 million dollar service contracting company generating 35% gross margin, the administrative applications in the business should be on the order of .08% of revenue, or about $8,000 per year on accounting, email, reporting, calendar, office productivity, etc.

If we total all of these up for a $10 million service contracting business, the percentage of revenue spent on technology applications is about 3.44% of revenue or about $344,000 per year. Now my ears are almost bleeding from the screams and bellows of “That’s Crazy!” that I can hear coming from service contracting customers reacting to this number. But is it so crazy? Are applications that help your business become competitive in attracting new customers, driving new revenue, and charging a premium price really worth that type of spending? Consider these two examples. How much do you pay for an application like Square that helps you collect money from a customer in the field? It consummates the sale by getting the cash now. You happily pay about 2.5% of revenue for this type of application. How about the central station monitoring application that enables you to sell a high margin monitoring service? You happily pay between 30% and 50% of revenue for this valuable addition to your service arsenal. So no, 3.44% of revenue is absolutely not crazy for a full set of applications that help you drive value in your business.