# Jordan Kanter -- Full Site Content

> Jordan Kanter is a Chicago-based AI, data, and product consultant with 20+ years of enterprise experience helping Fortune 500 brands turn data into outcomes.. This file contains the complete text of every page and post on jordankanter.com, concatenated as Markdown for large language models.

Generated: 2026-05-26. Canonical index: https://jordankanter.com/llms.txt

## Pages

### AI, Data, and Consulting Services

Source: https://jordankanter.com/services/  
Markdown: https://jordankanter.com/services.md

> "Jordan is incredibly innovative and a fantastic asset to any team" - Fortune 100 Executive

Most consulting engagements end the same way: a polished deliverable, an invoice tied
to hours, and a team  no better equipped than when the work started.

### How I work differently

- Priced on outcomes, not hours. You are buying a result, not my time.
- Accelerators that you keep. The tools and systems I bring in do not leave with me.
- Capabilities that transfer. Your team is more capable when I'm gone, not more dependent on me for answers.

### What I do

**AI strategy and implementation.** Most AI spend right now is subsidized by venture capital, not sustainable pricing. I help you identify the systems that are affordable and build the ones worth building.

**Data and intelligence.** Dashboards measure what's easy. The harder question is whether your data tells you something your competitors don't and whether you can act on it faster.

**Describe your ideal outcomes. I'll tell you if I can help deliver it.**

---

## Blog Posts

### AI and the README

Source: https://jordankanter.com/artificial-intelligence/2026/05/26/ai-and-the-readme.html  
Markdown: https://jordankanter.com/artificial-intelligence/2026/05/26/ai-and-the-readme.md  
Published: 2026-05-26  
Categories: artificial-intelligence

Most consultants do not understand the importance of a leave-behind now that AI is doing most of the coding:

Here is what changed in the last year and why the README or leave-behind playbook and documentation is more important than ever:

1. Code Understanding - AI Agents code faster then humans, but it is still incredibly important that humans understand the work.
2. AI Agent coordination - Multiple developers might have different agentic configurations. A streamlined way to coordinate between those agents requires documentation discipline.
3. Agent Economics - Token costs will balloon without constant vigilance on context information per token. Most AI systems have to read context fresh every time and therefore without constant compression and high information per token costs compound exponentially.

Ultimately, I have organizations save between 30-50% in token costs using my simple framework for context optimization.

---

### A Story About Organizations and AI

Source: https://jordankanter.com/artificial-intelligence/2026/05/19/a-story-about-organizations-and-ai.html  
Markdown: https://jordankanter.com/artificial-intelligence/2026/05/19/a-story-about-organizations-and-ai.md  
Published: 2026-05-19  
Categories: artificial-intelligence

Are you building for the AI future?

"AI-ready" operations don't mean the same thing to every company. I recently heard a story from a partner about two large organizations taking entirely different approaches to the future.

Org one wants SaaS-derived, "business-user ready" systems layered on top of AI . The org wants systems ready to execute cleanly, efficiently, and at scale across a well-governed matrix of capabilities. This org wants AI to improve every part of their business without changing their processes, and insists their vendors and dev partners shoulder the full cost of making their systems "enterprise ready."
No one in org one becomes an "AI builder."

Org two takes the opposite approach. Every individual, from the CEO on down, is given the training, skills, tools, frameworks, and charter to build automation around their specific domain expertise. Everyone in org two becomes an "AI builder."

Which organization is showing ROI on their AI investment?

Org two, and the contrast is striking.

Organizations that take org two's approach are showing
5–50% greater ROI and 10–20% lower operating costs from AI.
In this winner-take-all landscape, the org twos are winning.

Want to learn how to flex your "org two" muscle? Reach out for a [conversation](https://jordankanter.com/services#contact).

---

### LLMs and the wall of text

Source: https://jordankanter.com/artificial-intelligence/frontier-models/llms/2026/05/07/products-and-the-wall-of-text.html  
Markdown: https://jordankanter.com/artificial-intelligence/frontier-models/llms/2026/05/07/products-and-the-wall-of-text.md  
Published: 2026-05-07  
Categories: artificial-intelligence, frontier-models, llms

![The Wall Of Text](https://jordankanter.com/assets/images/wall_of_text.jpeg)

Product and engineering teams need to beware of the AI "wall of text". LLMs have a tendency to generate huge amounts of text even for simple answers. The is likely due to increased ability to reason coupled with the fact that we as humans tend speak ambiguously, resulting in a mirroring of such semantics on the part of the AI.

Developers have had decades to improve tooling for structured data, but only three years to improve tooling on the kind of unstructured data output by LLMs. Until harnesses and tooling catches up, here is how developers can avoid "wall of text syndrome"

1. Ask the model to "write at most 50 words before code" and "be brief".
2. Use shorter instead of longer words to describe the coding problem (the model will mirror this).
3. Establish standard practices around CLAUDE.md, skills, tools, and MCPs that reinforce brevity.

This isn't just best-practices, it is critical to reducing the amount of cognitive load on developers and reducing token costs while reducing time to value.

How are you dealing with the LLM wall of text in your organization?

---

### The Future of Consulting

Source: https://jordankanter.com/consulting/artificial-intelligence/frontier-models/llms/2026/05/07/how-consulting-will-change.html  
Markdown: https://jordankanter.com/consulting/artificial-intelligence/frontier-models/llms/2026/05/07/how-consulting-will-change.md  
Published: 2026-05-07  
Categories: consulting, artificial-intelligence, frontier-models, llms

#### What does the future of consulting look like in the age of AI?

There are many that call AI the death knell of consulting. Or that AI providers interest in
consulting is the obituary for the entire consulting industry.

I disagree. Rather, it is the obituary for a certain kind of consultany. The kind of consultancy that prints cash for "advice" or the consultancy that makes money using a mixture of cheap services and cheap labor. Neither of those types of consultancies survive the next five years. Unfortunately, that amounts to close to 70-80% of the industry.

But, the reason I stick to that 80% number (and not more) is counter-intuitive. And it has to do with the way AI companies are treating consulting.  AI companies are investing in the "Forward Deployed Engineer". Unfortunately, this is just a new name for a consultant who doesn't understand your business model.

Consulting teams will wake up to the fact that there is *massive* value in humans that can interact with AI, code, understand a client business model, and communicate effectively with client business stakeholders. Once they do they will adapt to the new reality and ultimately survive the "service-pocalypse"

[Learn more](https://jordankanter.com/services) about this new kind of consulting.

---

### Why the Semantic Web Failed

Source: https://jordankanter.com/data/2026/04/22/the-semantic-web-renaissance.html  
Markdown: https://jordankanter.com/data/2026/04/22/the-semantic-web-renaissance.md  
Published: 2026-04-22  
Categories: data

##### The Beginning

In 2004, the w3c came together to outline a way by which the web and semantics could be brought together. They called it the Semantic Web.

Unfortunately, what the w3c came to learn was the foundation by which one works with data semantics and the web are unfortunately orthogonal. Most people want the web to be simple,

somewhere they can hang out, a town square, something akin to Socrate's Agora. But, the semantic web was anything but simple. Learning to navigate the hybrid mesh of

RFCs, contradictory specifications, and complicated file formats was more than enough to drive an HTML author to Microsoft Word. However, a platform as rich as the web does not have to be

be the same thing to two people, and this ultimately is where I believe the semantic web failed.

Ultimately, the authors of Semantic Web specifications refused to hide the complexity of the underlying data structure, and left authors to make the massive leap from communicating

to authoring a graph and not just any graph, a graph with a massive amount of rules. The inference engine and the calculus behind it were not merited. What was merited was a product that introduced individuals

to the power of meaning to generate insights. In other words, data semantics were never for the everyman, they were never for the Marketer or

Blog author or for the social media user. The Semantic Web was always for the data nerd, and for a very specific kind of data nerd who was willing

and capable of reading the massive amount of specs (see [here](https://www.w3.org/TR), [here](https://whatwg.org), and [here](https://www.rfc-editor.org/rfc-index-100a.html) for a few),

and use the power of decentralization to make sense of it all.

In other words, the concept of the semantic web was good, leverage the power of the web as a decentralized data structure to drive crowd-sourced meaning

at scale. Take the massive amount of information that was getting pumped into the web in form of content and organize it intelligently,

and build underlying models that deeply understood the semantics of the information getting published.

##### The Good Parts

There was and is lots of good in treating the web as a semantic data source. First, it gets the run-of-the-mill web

developer thinking in terms of graphs in addition to applications. It allows for the richness of the web to be transferred to apis, and ultimately [Graphql](https://graphql.org)

was highly influenced by the Semantic Web. Applications like CWM suffered from this as well. They were so busy statically parsing that they
 lost the meaning of how mass audiences use the internet. This is why the semantic web utilities remain marginally popular despite some audiences having moved on to MCP and A2A as application platforms of choice

##### Why LLMs will usher in the Semantic Web Renaissance

What the Semantic Web was missing was a unifying "killer app" that could turn the graph into something more approachable. The LLMs is exactly that. In a sense, what ChatGPT and its Transformer brethren have accomplished is to turn semantics into an obviously useful system. It makes semantics something easy to interact with. It is what these data-starved LLMs require in order to be truly useful. Imagine for a second a world where explainable LLMs were a given, and we could give the models the context they needed without any effort. To me, this is the promise of the Semantic web. As a data engineer, I believe LLMs will show us why we (as data professionals) need the semantic web. We need it to inject existing semantics into our systems to give LLMs context and drive intelligent traversal of facts, figures, relationships, even information about the neural network itself.

---

### PRISM: A new Paradigm in Consumer Data

Source: https://jordankanter.com/data/artificial-intelligence/customer-experience/2026/04/22/prism-platforms.html  
Markdown: https://jordankanter.com/data/artificial-intelligence/customer-experience/2026/04/22/prism-platforms.md  
Published: 2026-04-22  
Categories: data, artificial-intelligence, customer-experience

There is a revolution happening in Customer Data. Ever since Tealium released AudienceStream and re-invented the way marketers interface with data, organizations have invested time and money in unifying customer data using CDPs (Customer Data Platforms) and CDP-like Customer 360 platforms.

As AI reshapes every aspect of business, it's time to introduce a new platform category that returns CDPs to their marketing activation roots while embracing their evolution into comprehensive customer service and insights platforms.

Introducing PRISM platforms: PRISM stands for Platform for Relationships, Insights, Service & Marketing. This is not a new kind of system. Rather, it is a convergence of three key industry trends. First, the explosive popularity of headless, warehouse-first CDP approaches (see RudderStack , Hightouch, and MetaRouter). Second, the growing integration between CDPs, CX, and Service platforms (see Lytics and Twilio Segment). Third, the trend of Marketing Automation Platforms toward intelligent data management (See Klaviyo for a great example).

What comes next? Organizations must prioritize AI-powered relationship building with customers and prospects, ensuring every technology investment directly supports clear business outcomes (regardless of a fancy new name for an old idea).

---

### 5 Signs Your Retail Customer Data Strategy Needs an Overhaul

Source: https://jordankanter.com/retail/customer-experience/data/2026/04/22/5-signs-your-retail-customer-data-strategy-needs-an-overhaul.html  
Markdown: https://jordankanter.com/retail/customer-experience/data/2026/04/22/5-signs-your-retail-customer-data-strategy-needs-an-overhaul.md  
Published: 2026-04-22  
Categories: retail, customer-experience, data  
Tags: blogpost

In today's hyper-competitive retail landscape, your customer data strategy isn't just a technical consideration—it's the backbone of your competitive advantage. With digital-native retailers setting new standards for personalized experiences and major players investing heavily in AI-powered customer insights, many traditional retailers are finding themselves at a critical crossroads.

But how do you know if your current approach to customer data is holding you back? Here are five unmistakable warning signs that your retail customer data strategy needs a comprehensive overhaul:

#### 1. Your Teams Are Working with Different Customer Counts

If your marketing team reports one customer count, your loyalty program another, and your e-commerce platform yet a different figure, you're looking at a fundamental data integrity problem. This discrepancy isn't just an administrative headache—it undermines your ability to make strategic decisions and accurately measure performance.

When different departments can't agree on basic customer metrics, it's impossible to align on measures like customer lifetime value or churn prediction. This fragmentation creates a ripple effect that impacts everything from inventory planning to marketing budget allocation.

#### 2. You Can't Recognize the Same Customer Across Channels

Today's shoppers move fluidly between your physical stores, website, mobile app, and social media channels. If your systems treat a customer who shops in-store and then online as two different people, you're missing crucial opportunities for personalization and likely frustrating your customers.

Consider this scenario: A customer researches a product on your website, visits your store to try it, then completes the purchase on your mobile app. If you can't connect these touchpoints, you might bombard them with ads for a product they've already purchased or miss the chance to recommend complementary items based on their full purchase history.

#### 3. Your Personalization Efforts Feel Generic to Customers

"Hello [FIRST_NAME]" isn't personalization in 2025—it's the bare minimum. If your "personalized" recommendations regularly include products that are irrelevant or items a customer has already purchased, your data foundation isn't supporting true personalization.

Advanced retailers are delivering experiences where product recommendations, pricing strategies, and even in-store experiences are tailored to individual preferences and behaviors. If your personalization still feels like mass marketing with a name attached, your customer data strategy isn't enabling the experiences today's consumers expect.

#### 4. You Can't Answer Basic Questions About Customer Behavior

When leadership asks questions like "What's the average time between first purchase and second purchase?" or "Which product categories tend to drive the highest customer retention?" your team should be able to provide answers quickly. If these types of questions require custom analytics projects or weeks of data preparation, your customer data strategy isn't delivering actionable insights.

The most successful retailers can seamlessly answer complex questions about customer segments, purchase patterns, and cross-channel behavior. This capability isn't just about having the right technology—it's about having a data strategy that makes customer insights accessible throughout the organization.

#### 5. Your AI and Advanced Analytics Initiatives Keep Stalling

If your organization has invested in AI, machine learning, or advanced analytics projects that consistently underperform or never make it to production, your customer data foundation may be the culprit. These technologies rely on clean, unified, and accessible customer data to deliver meaningful results.

Many retailers find themselves in a frustrating cycle: investing in sophisticated AI capabilities only to discover their underlying data isn't ready to support them. This leads to expensive projects that deliver disappointing ROI and growing skepticism about data initiatives across the organization.

#### The Path Forward

Recognizing these warning signs is the first step toward building a customer data strategy that can power truly competitive retail experiences. The good news is that overcoming these challenges doesn't necessarily require starting from scratch or making enormous technology investments all at once.

The most successful retailers are taking a pragmatic, phased approach—starting with creating a unified customer identity across channels, then progressively enhancing their capabilities for personalization, prediction, and automation. With each step, they're not just improving their technology but evolving their organizational approach to becoming truly data-driven.

As we move deeper into the AI era of retail, the gap between leaders and laggards in customer data capabilities will only widen. The retailers who recognize these warning signs and take decisive action to address them will be positioned to thrive, while those who maintain the status quo risk falling permanently behind.

_Is your retail organization showing any of these warning signs? We'd love to hear about your experiences and challenges with customer data in the comments below._

---

### AI and Data: The state of the Union 2025

Source: https://jordankanter.com/artificial-intelligence/machine-learning/data/2025/11/05/ai-and-data.html  
Markdown: https://jordankanter.com/artificial-intelligence/machine-learning/data/2025/11/05/ai-and-data.md  
Published: 2025-11-05  
Categories: artificial-intelligence, machine-learning, data

### AI and Data : The State of the Union 2025

The world is cyclical. We're in a new phase of AI which will be dominated by the power of data, but not in necessarily in
the way that most think. In reality, many see the state of AI as "hype" or a "bubble".

---

### Journey Science : How Journey Science is Transforming Data Science and Customer Experience

Source: https://jordankanter.com/data/customer-experience/marketing/artificial-intelligence/2025/09/16/journey-science.html  
Markdown: https://jordankanter.com/data/customer-experience/marketing/artificial-intelligence/2025/09/16/journey-science.md  
Published: 2025-09-16  
Categories: data, customer-experience, marketing, artificial-intelligence

#### The beginning

In the beginning, Computer Science was abstract. The theoretical concepts of Boole, Turing, Babbage, and Lovelace quickly gave way to real-life innovations. Their focus was ambitious but simple: Change the world by making our lives more efficient and less error-prone, relying on machines to do repetitive tasks humans were less suited to. This would let humans focus on functions that mattered more, like advancing science or the arts.

#### Data Science: The New Evolution

Almost immediately, computers were leveraged for statistical analysis. From the first spreadsheet application (Visicalc on the Apple II) to Fortran to Peter Naur’s “data science,” the history of computers and data are intimately intertwined. Like Computer Science, the era of paper spreadsheets and hand-written calculations for Data Science gave way to parallel machines with the ability to crunch thousands upon thousands of metrics on special-purpose chips (GPUs).

Next, AI and Machine-Learning became critical capabilities for the Enterprise. With Machine Learning, computers immediately became proficient in expertise and traditional human skillsets such as tagging images with what they depict or recognizing the words spoken in an audio clip. Similarly, Customer Experience professionals looking to uplevel consumer engagement and redefine what’s possible looked to AI and Machine Learning to build new experiences. Today, CX and UX teams create predictive, adaptive, and intuitive experiences that act in ways that delight and amaze customers. While this phenomenon transcends the discipline of CX, seeing the impact of AI on CX is as easy as looking at the “recommended products” on Amazon’s home page. Like Amazon, forward-looking retail brands provide virtual shopping experiences that listen and engage like a real customer service agent, reacting to choices and predicting future tastes based on trends, transactions, and preferences.

#### Journey Science: The Next Innovation
As such forward-looking organizations leverage AI to compete, Customer Experience teams are uniting Computer Science and Data Science into a holistic discipline and methodology. Called Journey Science, these teams manage journeys and personas via evidence, orchestrating new actions that adapt and predict what will resonate with any customer. Similarly, Journey Science practitioners apply Computer Science and Data Science lessons to drive 1-to-1 personalized experiences tailored to every consumer at every touchpoint. By evaluating ROI at every step, Journey Science practitioners adapt to trends in real-time and more quickly than competitors, allowing them to quickly build experiences crafted to “always online” and “digital-only” customers. Last, by adopting empirical and rigorous metrics, Journey Science practitioners monitor, design, and predict how to build the next generation of impactful omnichannel experiences.

---

### Measuring New User Interactions Types: How Touch is Changing Marketing

Source: https://jordankanter.com/strategy/interaction/analysis/marketing/customer-experience/2025/09/16/analyzing-new-interaction-models.html  
Markdown: https://jordankanter.com/strategy/interaction/analysis/marketing/customer-experience/2025/09/16/analyzing-new-interaction-models.md  
Published: 2025-09-16  
Categories: strategy, interaction, analysis, marketing, customer-experience

##### The Why

Customers are beginning to interact with brands in new and interesting ways. Clicks are giving way to touch, and touch is giving way to eye movements, voice, and biometrics.
In a 2018 [eMarketer survey](http://totalaccess.emarketer.com/chart.aspx?r=217562), 'voice interaction', 'multidevice interaction', 'Virtual reality/Augmented reality',
'Wearable Tech', and 'Gesture-based interaction' were among the most perceived impactful visible marketing trends,
with greater than 100% of respondents believing one of those trends will affect Customer Experience worldwide.
Therefore, understanding these new interaction modalities and how to leverage them is critical to brand success.

While marketers expect these technologies to impact results, teams have struggled to leverage these new technologies.
The reason is twofold. First, specialty hardware and software is required to build such experiences. Second, software may be difficult, hard-to-use, or not designed for marketer. Therefore, make sure to experiment with platforms to understand the team's comfort with the technology. Thorough experimentation will enable your team to learn, as well as begin to expose new insights immediately.

Thorough experimentation will also generate new ideas for elucidating such behavior.
For example, a VR racing game can embed branded visuals or interactive content embedded in gameplay, such as virtual billboards or custom car paint jobs. Measuring eye-tracking movements or gestures enabled by those interactions could answer questions like "What colors did the user select in the color-slider before settling on the final color?" or "did the user look at the branded content and if so, how long?".

##### The How

From a metrics standpoint, such metrics can provide better viewability for ads.
Using the above racing example, rather than typical PPC or PPM pricing structures,
a marketing department could negotiate a rate based on 'pay per seconds of gaze',
in which the purchaser only pays when the user looks at the branded visual on screen.
One could easily imagine these interactions broken down into 'quick scan' (looks under .5 second), by 'eye' (right or left), as well as 'attenuated glance' (any look with both eyes over a particular threshold). Such interactions tell a different story than a simple click, hover, or scroll.

While eye-tracking metrics represent a fascinating ideal for advertisers, behaviors such as touch
represent unique and no less rich sources of data for brands. New constructs such as 'tap', 'swipe', 'multi-touch' (touch of interface with more than one digit), 'pinch', and 'twist', represent fascinating insights and tell much richer stories on their own than a click, and a set of these interactions tell a much different story about the customer experience than a series of clicks.

One options is to score the alternate gestures along with clicks, generating aggregate scores and with each of the above interactions adding points to the aggregate score. For example, a 'click' or a 'tap' worth 1, a pinch worth 2, and a 'twist' worth 7. Thus, the score for a screen in which the user clicked 4 times would be equal to two pinches on the same screen. Add in conversion funnel optimization and it becomes increasingly clear that these new interactions are often a lost piece of feedback to brands, and generating such metrics will be critical to future brand success.

---

### JavaScript as a Strategy

Source: https://jordankanter.com/javascript/programming/2025/09/16/javascript-as-a-strategy.html  
Markdown: https://jordankanter.com/javascript/programming/2025/09/16/javascript-as-a-strategy.md  
Published: 2025-09-16  
Categories: javascript, programming

let MachineLearning = new ReinforcementLearner()

This afternoon, I stumbled upon Jeff Atwood's ["Principle of Least Power"](https://blog.codinghorror.com/the-principle-of-least-power/), and it occurred to me that
perhaps some of the machine learning strategies I have been helping clients with could be implemented in javascript.

##### Why Javascript?

The short answer is popularity, and the long answer is flexibility, speed, and short learning curve (I guess it is not so long).
Which leads me to the idea that

##### Javascript is a Strategy

In the near future, coding will be a requisite skill for employees. But choosing a first programming language can be tough,
in that understanding a programming languages strength and weaknesses are typically topics for more seasoned programmers.
Therefore, I recommend you find a project that interests you, and the learn the language that people working on those projects
are learning. For example:

- Interested in the web or building something visual? [Javascript](https://developer.mozilla.org/en-US/docs/Web/JavaScript) is a great first choice.
- Interested in data analysis? Maybe try [Python](https://python.org), then [SQL](https://en.wikipedia.org/wiki/SQL), then [R](https://www.r-project.org/). Python,
being designed for ease of learning, makes for a great first language.
- Interested in high-performance computing or computer architecture? Try C++.
- Interested in distributed computing? Try [Elixir](https://www.elixir-lang.org)
- Interested in symbolic computation or the study of computer languages themselves? Try [LISP](http://common-lisp.net) or [Scheme](http://www.schemers.org)

and so on...

But, then why do I say Javascript is a strategy if you can do so much with so many different languages?
It turns out that everything *but* writing code is a hassle. Installing, configuring, and getting packages
for different languages tends to suck up time. But, good news, [you are most likely set up to code with javascript already](https://medium.com/@isaaclyman/when-you-finish-reading-this-youll-know-how-to-code-721339942b51)

Happy coding.

follow me [@jikanter](https://twitter.com/intent/follow?screen_name=jikanter) for more.

---

### How to find the perfect balance between Marketing, Sales, and Customer Service

Source: https://jordankanter.com/product/marketing/customer-experience/2025/09/16/how-to-find-the-perfect-balance-between-marketing-sales-and-customer-service.html  
Markdown: https://jordankanter.com/product/marketing/customer-experience/2025/09/16/how-to-find-the-perfect-balance-between-marketing-sales-and-customer-service.md  
Published: 2025-09-16  
Categories: product, marketing, customer-experience

#### Why find a balance?

Even a perfect product needs sales and Marketing. Contrary to what some in Silicon Valley would have you believe, products do not sell themselves. Every product needs the following:

1. Integrated and Omni-Channel Marketing to educate the public about your product and/or service
2. Sales and Account Management to deliver your organizations story and negotiate deals
3. A robust data-driven evaluation (analytics) on how Sales and Marketing are performing at the above two efforts.

As someone who works in sales and marketing, I can tell you that about 90% of businesses do not measure enough or measure the right metrics.

But what about the other 10%? The 10% who obsess about sales and marketing metrics and fight tooth and nail for every conversion? The 10% who put an overlay on every page with an offer, the 10% that send unsolicited mail or emails, calling phones at every hour, or aggressively buying digital ad space to get every ad on every site you browse showing you their product. This is the same company that has sales and account management up at all hours of the night updating thousands of fields manually in salesforce after a call, and obsessively tracks every dollar spent in favor of those sweet, sweet conversion.

This company is a data-driven marketer’s dream, right? This is the type of company that gets it, right?

#### Enter Customer Experience (CX) Metrics

For that 10%, What if conversion matters less then the organization thinks? After decades of consulting, examining the data of organizations large and small, I have noticed a pattern that transcends organizational size, sector, and model. This pervasive pattern is that the 10% of organizations with outsized investment in Sales, Marketing, and metrics therein experience diminishing conversion returns once they hit 6%  lift. I’ll say it again for the kids in the back: **After 6% improvement metrics are reached, investing more money in sales and marketing experimentation (a/b testing and conversion optimization) is a bad idea (in a vacuum).**

To those not millennials or younger, this makes no sense. How could that be? Traditionally (say before the year 2000), a CEO investing in Sales and Marketing could consider that money well spent.

But as my colleague Brian Flanagan likes to say, today’s customer touchpoints are “a complex pinball of events” some which are unintuitive, hidden, or post-conversion.

But what about that white whale metric? The Wonka’s golden ticket metric that puts great businesses on top and keeps them on top? Before I get there, I will propose that all good metrics have a foundational OKR or strategy that the metric attempts to measure. And that magic OKR is:

Deliver a great product or service that your customers love, and provide industry leading customer service atop said industry leading product.

In other words “Deliver Excellence”.

And the corresponding metric that indicates the business is indeed delivering excellence:

Perficient Platinum Performance(tm) Score = (Customer Satisfaction Score) \* (Customer Lifetime Value)

For those who have been working in Digital Analytics or Customer Data for a bit, the above idea is anything but earth shattering. But for Growth hackers, product marketers, and old-school CEOs chasing vanity metrics, the above metric is a north star needed to

This number tells business leaders how well their business is performing without having to worry about confounding variables. In fact, I would argue any business more focused on net new revenue or cost reduction or sales efficiency or Multi-Channel Attribution or whatever predictive potion the data science team cooks up in favor of the above score is doomed to fail.

Let’s break down the Perficient Platinum Performance (PPP) score:

Customer Satisfaction Score (CSAT): A value, between 1 and 10 and defaulting to zero (in case a customer has never reported their satisfaction) that is measured by asking a customer (either via digital survey or directly) about their experience. This number can be NPS or some other metric, but it always starts with an **explicitly reported value** and then blends more data in from there

For example, it could include a metric of willingness to try new features or complete in-product or post-purchase offers. It could include some level of content consumption or affinity and sentimentality metric. But at the end of the day the following does just fine as a starting point:

“On a scale of 1 to 10 (from 1 being never to 10 being incredibly likely)  please rate how likely you are to recommend  product and services to a friend”

Or

“On a scale of 1 to 10 (1 being awful and 10 being the most incredible experience) please rate your overall satisfaction with your service from  as it relates to your recent purchase”

Customer Lifetime Value (CLTV):

The sum total of all transactions over the entirety of the customers lifetime.

This sub-metric is simple. Take all the money the individual has spent on your organization and sum it up.

#### The interplay between Marketing/Sales and CX

Now, armed with the Perficient Platinum Performance (PPP) score, we can see clearly that see Marketing and Sales has a critical place in overall organizational performance. But optimizing those activities should never detract from the overall customer experience. And that is exactly what those 10% of businesses discussed above are doing. Just as an incredible product will never sell itself, no amount of intelligent Marketing or Sales will sell a bad product or service. In fact, by overdoing the Marketing bubbles or Sales emails, organizations can hurt the customer experience - a fact which is reflected in the PPP score. While those individuals may buy once, the data I have examined shows those individuals are neither satisfied nor repeat customers.

Ultimately, the way to growth is to win Customer’s hearts. And the way to win customers’ hearts is to provide an excellent product with exemplary customer service.

Interested in learning more? Feel free to reach out on LinkedIn or via the form below!

In follow-ups, I will deep dive into unintuitive sub-metrics that matter, how Sales and Marketing practitioners of Journey Science can influence the Perficient Platinum Performance score, and how to measure the ROI of every activity using Journey Science and the Perficient Platinum Performance score.

---

### Artificial Intelligence - Rules of Engagement and the Customer Journey

Source: https://jordankanter.com/product/strategy/artificial-intelligence/machine-learning/customer-experience/2018/02/20/artificial-intelligence.html  
Markdown: https://jordankanter.com/product/strategy/artificial-intelligence/machine-learning/customer-experience/2018/02/20/artificial-intelligence.md  
Published: 2018-02-20  
Categories: product, strategy, artificial-intelligence, machine-learning, customer-experience

AI has changed how we relate to the customer and the customer journey.
Specifically, deep machine learning will impact the landscape in the next year.
In 2018, I suspect forward-looking marketers will begin to harness AI to engage larger and more targeted audiences.
Decisions on AI will allow marketers to develop better campaigns delight our customers, and as such earn repeat business and drive brand loyalty.
Therefore, understanding and leveraging AI is critical to a brand's success in 2018.

#### What is AI?

In general, AI is simple. Rather than giving our software the set of rules to execute tasks, we provide fuzzier rules by which it must abide, and run our software in this "world". In other
words, rather than providing strict input to our software, we allow our software to acquire input from the 'world', and then derive from that input how it could and should behave. To better illustrate,
imagine trying to tell some software how to "make a peanut butter and jelly sandwich". We would provide rules to take two slices of bread, take a knife out of the drawer, spread the peanut butter and jelly, and place the two slices together.
In contrast, if we were to "teach" an AI to do the same task, we might provide some basic functionality explicitly, but attempt to "teach" the system to make the sandwich by abiding by a certain set of generic rules, as well as an
'understanding' of how the system should calculate responses to input. It is those calculations that provide the 'Intelligence' to the system. In other words, it is these calculations that give the system the ability to actually
make the sandwich.

##### How do I use it as a marketer?

Today, the most forward-thinking Companies rely on AI to optimize marketing initiatives. For example, companies use AI to automatically bid on ad-space, learning hidden rules to optimize visibility.
Companies use AI to better serve customers, driving struggling customers to customer service representatives or online chat. Companies use AI to run and design optimizations, developing new features by allowing software to choose when and how to run experiments.
Companies use AI to sift through mountains of data, automatically generating visuals, prose, and analyses. Companies can use AI to decide the content and timing of campaigns and given historical data can use AI to decide
what type of campaigns will generate the most lift in the future.

##### Should I use AI?

While AI is powerful, organizations should not invest in AI to solve all problems. While simpler problems with straight-forward rules can be solved with AI, the complexity of constructing AI increases costs, and simpler problems do not merit such an investment.
To go back to the peanut butter and jelly sandwich, teaching an AI such a simple task could be an immense undertaking, and therefore organizations would be wise to simply input the rules themselves rather than utilize AI.

##### Responsible AI

AI is powerful. Like any powerful technology, the potential to misuse AI should give organizations pause before employing them to solve problems.
Some of the risks include malevolent modifications of existing AI systems, job displacement in favor of AI, and privacy concerns as a result of the large number of data sets input into an AI. Organizations would do well to engage an expert to design, assess, construct, measure, and evaluate their AI strategy.

---