ascend-growth-strategies-d1rth

CFOs Are Moving on AI. Are You?

Fri, 14 Nov 2025 16:24:47 GMT

Practical insights on building literacy, governance and experimentation.

The body content of your post goes here. To edit this text, click on it and delete this default text and start typing your own or paste your own from a different source.

The Future of Service

Wed, 22 Oct 2025 17:52:18 GMT

Building a Human AI Symbiosis

The body content of your post goes here. To edit this text, click on it and delete this default text and start typing your own or paste your own from a different source.

RAG vs. CAG

Wed, 22 Oct 2025 17:49:52 GMT

A Decision Framework for Business Leaders

The body content of your post goes here. To edit this text, click on it and delete this default text and start typing your own or paste your own from a different source.

Retraining is a Necessity

Wed, 22 Oct 2025 17:47:11 GMT

Every deployed AI model has an expiration date. The only question is: how fast?

Deploying a machine learning model is not a "set it and forget it" operation. The data that a model was trained on represents a snapshot of the world at a specific time. As the world changes, the model's understanding becomes outdated, leading to performance degradation. Without periodic retraining, even your best models become liabilities.

Why Retraining is Essential

Data Drift – The most common culprit. The characteristics of incoming data change over time.

Example: A demand forecasting model trained on 2023 seasonal trends might become unreliable if consumer purchasing habits shift dramatically in 2025 due to new competition, economic recession, or supply chain disruptions resulting in inventory misallocations costing millions in markdowns and lost sales.
Impact: The input data the model sees in production starts to look different from what it was trained on, causing predictions to drift from reality.

Concept Drift – The relationships between input data and the target variable (what you are trying to predict) change.

Example: A fraud detection model learns what "fraud" looks like based on historical patterns. However, fraudsters constantly evolve their tactics. What was a clear indicator of fraud last year might not be one today, or entirely new fraud schemes emerge that the model has never encountered. A credit card fraud model that is not retrained can see false negative rates double within 6-12 months as criminals adapt.
Impact: The underlying rules the model learned are no longer accurate for predicting outcomes, leading to increased fraud losses or customer friction from false positives.

Model Degradation – Over time, without retraining, a model's predictive accuracy will certainly decline. Think of it like an athlete whose performance deteriorates without consistent training and adaptation.

Impact: Gradual erosion of ROI as the model's effectiveness diminishes, often slowly enough that teams don't notice until significant value has been lost.

New Information / Features – As your business evolves, you gain access to new data sources or identify new features that are highly predictive. Customer behavior data, competitive intelligence, or operational metrics that did not exist when the model was first built can dramatically improve performance.

Impact: Competitors who leverage newer data sources will outperform your static models, putting you at a strategic disadvantage.

Addressing Bias – If initial training data contained biases, or new biases emerge in real-world data, retraining with more diverse and balanced datasets can help mitigate these issues and reduce regulatory risk.

Impact: Unchecked bias can lead to discriminatory outcomes, regulatory penalties, reputational damage, and loss of customer trust.

The Cost of Inaction

Failing to retrain models is not just a technical oversight, it's a business risk with measurable consequences:

Revenue Leakage: Inaccurate predictions lead to missed opportunities, from lost sales to suboptimal pricing decisions.
Increased Operational Costs: False positives create unnecessary work; false negatives let problems slip through undetected.
Regulatory and Compliance Risks: Models that perpetuate bias or fail to meet performance standards can trigger regulatory scrutiny and fines.
Competitive Disadvantage: While your models decay, competitors with active retraining programs pull ahead in accuracy and customer experience.
Eroded Trust: When stakeholders lose confidence in model outputs, they revert to manual processes or gut instinct, undermining your entire AI investment.

The difference between a thriving AI program and a failed one often comes down to model maintenance discipline.

When to Retrain Models

The frequency of retraining depends heavily on the specific use case and how dynamic your data environment is.

Scheduled Retraining: For many applications, models are retrained on a regular cadence (daily, weekly, monthly, or quarterly). This works well when data changes predictably or when continuous monitoring is not feasible.
Performance-Based Retraining (Triggered): Often the optimal approach. Monitor the model's performance continuously in production. When accuracy, precision, recall, or other key metrics drop below a predefined threshold, automatically trigger a retraining process. This ensures you are responding to actual performance degradation, not arbitrary timelines.
Data Drift-Based Retraining: Monitor the characteristics of incoming data itself. If there is a significant shift in the distribution of input features compared to training data, trigger retraining before performance visibly degrades.
Business Event-Based Retraining: Major business changes such as launching a new product line, expanding to new markets, executing a significant marketing campaign, or responding to competitive moves may necessitate immediate retraining to capture new patterns.

The Bottom Line

The question is not whether to retrain your models, it is whether you have the systems to know when and the processes to do it right.

Organizations that treat model maintenance as an operational discipline, not an afterthought, turn AI from a depreciating asset into a compounding competitive advantage. They build monitoring infrastructure, establish clear performance thresholds, and create automated pipelines that make retraining routine rather than reactive.

In today's rapidly changing business environment, your models need to evolve as fast as your markets do. The companies that win are those that recognize AI operations as a continuous capability, not a one-time implementation.

Is your organization prepared to keep its models current or are you flying blind with yesterday's intelligence?

The AI Compliance Gap

Wed, 22 Oct 2025 17:41:58 GMT

Why Your Most Critical Question has No Verifiable Answer

Last month, three major AI providers (OpenAI, Google, and Anthropic) quietly reversed user privacy protections, transforming consumer conversations into permanent training data by default. While enterprise customers remain protected by different policies, this mass policy shift exposes a deeper problem that should concern every executive evaluating AI: when vendors make claims about their AI systems, there is no independent verification framework to prove those claims are true.

For business leaders accustomed to the rigorous compliance infrastructure of financial services, where SAS 70 evolved into SOC 2 and every control can be audited, the AI industry's approach is jarring. You cannot independently verify what data trained the AI you are about to deploy. You cannot audit the composition of training datasets. You cannot get third-party attestation about data sources. You are expected to trust vendor claims without the verification mechanisms that have been standard practice in enterprise technology for decades.

This isn't a technical problem. It is a governance crisis hiding in plain sight.

The Foundation You Can't Inspect

Every AI model is a direct reflection of its training data. Quality data produces quality outputs. Biased data produces biased outputs. Contaminated data produces unreliable outputs. This relationship is absolute and unavoidable.

Modern large language models are trained on massive datasets, often hundreds of billions or even trillions of words. The process requires thousands of human annotators to label examples, classify content, and provide the evidence models need to learn appropriate behavior. Want the model to refuse bomb-making instructions? You need labeled examples. Want it to avoid racist outputs? You need curated training data demonstrating appropriate responses.

The challenge is not that this data exists. The challenge is that its specific composition remains one of the most closely guarded secrets in the tech industry.

Ask your potential AI vendor these questions:

What percentage of your training data came from social media versus curated sources?
How do you handle copyrighted content in training datasets?
What specific steps filter out harmful or biased content?
Which data sources were explicitly excluded and why?

The answers, if you receive them at all, will be vague references to "web crawls," "publicly available data," and "rigorous filtering processes." No specifics. No audit trail. No independent verification.

Why Opacity Persists

Major AI providers cite three reasons for training data secrecy:

Competitive advantage: Revealing data sources would help competitors replicate their models. This is defensible. No company wants to hand its secret sauce to rivals.
Legal complexity: Many models were trained on data of questionable provenance, including copyrighted materials now subject to multiple lawsuits. Disclosure could increase legal exposure.
Cost and scale: Properly documenting the provenance of petabytes of training data is expensive and time-consuming. When you are racing to ship products, documentation becomes a "later" problem.

But consider this: if your company were procuring any other enterprise-grade system (ERP software, a trading platform, a medical device), you would demand complete transparency about its components, testing methodology, and failure modes. Why should AI be different?

The answer is simple. It should not be. But unlike other enterprise systems, AI currently lacks the compliance infrastructure to make transparency verifiable.

From Opacity to Risk: The Business Impact

Deploying AI without understanding its training data introduces specific, measurable business risks:

1. Reputational Damage

Large language models generate statistically probable text sequences based on training patterns, not logical deductions from verified facts. This fundamental characteristic means they will generate false information: inventing facts, misrepresenting policies, or interacting with customers in brand-damaging ways. An unvetted model trained on unknown data sources is a reputational risk you can't quantify because you do not know what's in it.

2. Legal and Compliance Exposure

If your AI was trained on copyrighted content, biased datasets, or scraped personal information, you may inherit liability even if you did not create the model. The current wave of copyright lawsuits against AI providers, filed by authors, publishers, and music companies, demonstrates this risk is real, not theoretical.

3. Unpredictable Performance

General-purpose models trained on broad web data may perform inconsistently on specialized tasks. Without knowing what domain-specific content exists in training data, you cannot predict where the model will excel or fail. For customer-facing applications where accuracy is non-negotiable, this unpredictability is unacceptable.

4. Data Sovereignty Concerns

While major providers now contractually guarantee that enterprise customer data will not be used for training (verified through SOC 2 attestations of access controls), no framework exists to verify what data sources were used during the model's foundational training. You are taking vendor claims on faith.

The Compliance Infrastructure Gap

Here's where AI diverges sharply from mature enterprise technologies.

In financial services, the path from SAS 70 to SOC 2 created standardized attestation frameworks. Third-party auditors verify controls, validate claims, and provide independent assurance. When a vendor says they follow specific data handling procedures, you can demand proof through established audit frameworks.

In AI, no equivalent exists.

What Current Frameworks Cover

SOC 2 Type 2 attestations can verify:

Access controls are properly implemented.
Audit logs track who accessed customer data.
Data retention policies are followed.
Security procedures meet standards.

ISO 42001 (published December 2023) addresses:

AI governance processes.
Risk management frameworks.
Ethical AI practices.
Accountability in AI operations.

ISO 27001, FedRAMP, and other standards provide additional security and compliance verification.

What They Cannot Verify

None of these frameworks address the foundational question: What data trained this model?

They do not verify:

Training data composition or sources.
Whether specific content types were included or excluded.
Data provenance for the massive pre-training datasets.
Claims about filtering or curation processes.
Historical training decisions made before the model reached market.

This is the gap. Unlike financial auditing where you can trace every transaction and validate every calculation, AI training data exists in a compliance blind spot. Vendors make claims. You trust them. No third-party verification mechanism exists to confirm those claims are accurate.

For a business leader used to "trust but verify," AI offers only "trust."

Taking Back Control: Strategies for Smarter AI Integration

The solution is not to abandon AI. It is to approach deployment with the same strategic rigor you'd apply to any core business transformation. Instead of accepting a massive, opaque model as-is, implement controlled and transparent solutions tailored to your environment.

Solution 1: Retrieval-Augmented Generation (RAG)

For use cases demanding high accuracy, combine the language model with a vector database of your verified documents. The model retrieves information from your knowledge base before formulating responses, dramatically reducing hallucination risk and grounding outputs in verified facts. RAG provides the factual control that general-purpose models lack.

Best for: Customer support, internal knowledge bases, compliance-sensitive applications.

Cost consideration: Moderate initial setup, lower ongoing training costs.

Verification: You control the knowledge base content.

Solution 2: Fine-Tuning on Curated Data

Refine an off-the-shelf model using your company's specific data corpus. This aligns outputs with your knowledge base and brand voice, reducing the risk of inappropriate responses drawn from unknown training sources.

Best for: Domain-specific applications, branded interactions.

Cost consideration: Higher than RAG, requires technical expertise.

Verification: You control the fine-tuning dataset (but not the base model).

Solution 3: Smaller, Purpose-Built Models

The highest ROI often comes from specialized models trained on high-quality, domain-specific data rather than massive general-purpose models. These can provide more accurate, efficient, and safer results for targeted business functions.

Best for: Specific, well-defined tasks with clear success metrics.

Cost consideration: Potentially highest upfront, but purpose-built for your needs.

Verification: Maximum control over training process and data.

The Cost-Benefit Reality

These approaches require investment. Fine-tuning demands ML expertise. Purpose-built models require even more resources. For many mid-sized organizations, a well-implemented RAG system using a general-purpose model may offer the best balance of cost, risk mitigation, and performance.

The key is matching your approach to:

Use case criticality
Risk tolerance
Available budget and expertise
Performance requirements
Compliance obligations

A Fortune 500 company handling sensitive financial data should implement comprehensive controls. A startup using AI for internal brainstorming might accept more risk in exchange for lower costs. The mistake is treating all use cases identically.

Your Due Diligence Checklist

Before committing to an AI vendor, demand clear answers to these questions. If you receive vague responses or outright refusals, that tells you something important about the vendor's approach to transparency.

Training Data Provenance

How was the model trained, and what is the general composition of training data?
What guardrails were in place during data collection and curation?
Can you provide examples of data sources explicitly excluded and why?

Bias and Safety

What harmful or biased content existed in training data, and what specific steps mitigated its influence?
How do you test for and address bias in model outputs?
What ongoing monitoring detects emergent biases or safety issues?

Output Control

How can we ensure the model doesn't generate responses violating our policies or damaging our brand?
What controls exist to prevent hallucinations in our specific use case?
Can we implement hard guardrails for unacceptable outputs?

Model Updates and Versioning

How often is the model retrained or updated?
What's the process for incorporating new data?
How much advance notice do we receive about model changes that might affect our application?

Data Sovereignty and Security

What contractual guarantees protect our proprietary data from being used in future training?
Can you provide SOC 2 Type 2 or ISO 42001 attestations?
What audit rights do we have to verify compliance?
Where is data physically stored, and who has access?

Compliance Framework

What independent audits verify your claims about data handling?
Can we review audit reports under NDA?
What happens if training data sources become subject to legal challenges?
Do you indemnify customers for copyright claims arising from model outputs?

Customization and Control

If we fine-tune with our data, what does that process look like?
How is our data segregated during fine-tuning?
Can we implement RAG with our knowledge base?
What visibility do we have into model decision-making?

Critical note: Many vendors will provide only partial answers, citing competitive concerns. That is their right. But insufficient answers should factor heavily into your risk assessment and decision-making. A vendor unwilling to provide transparency about their systems is asking you to accept unquantifiable risk.

The Path Forward

The age of AI is here, but the era of blind trust must end. The compliance infrastructure that makes other enterprise technologies auditable and verifiable simply does not exist yet for AI training data. While standards like SOC 2 and ISO 42001 verify security controls and governance processes, they do not (and currently can't) verify the foundational question: what's actually in the training data?

This gap will not persist forever. As AI moves from experimental technology to business-critical infrastructure, demand for independent verification will drive the creation of new attestation frameworks. Organizations like NIST, ISO, and industry consortiums are working on AI-specific standards. But today, that infrastructure is nascent at best.

Until then, you have two choices:

Accept the opacity and deploy general-purpose models, managing risk through careful use case selection, robust testing, and accepting some level of unquantifiable risk.
Reduce your dependency on opaque foundational models through RAG, fine-tuning, or purpose-built solutions where you control more of the data pipeline.

Most organizations will need a portfolio approach: using different strategies for different use cases based on risk tolerance, budget, and technical capability.

True competitive advantage won't belong to companies that adopt AI the fastest. It will belong to those that adopt it the smartest, with clear-eyed assessment of what they know, what they do not know, and what they can verify. In the absence of mature compliance infrastructure, this requires asking harder questions, demanding better answers, and building verification mechanisms into your procurement and deployment processes.

The technology is powerful. The risks are real. And what you do not know about the training data will impact your bottom line, whether through reputational damage, legal liability, operational failures, or competitive disadvantage from poor model performance.

In business, trust has always required verification. AI should be no exception. The industry will eventually build the compliance infrastructure to make verification possible. The question is: will you wait for that infrastructure, or will you build your own verification mechanisms in the meantime?

Your competitors are making that choice right now. Make sure you are making it consciously, not by default.

Is Leadership Focused on the Right Kind of AI?

Fri, 12 Sep 2025 21:50:44 GMT

The Right AI will Result in ROI for you Business

While Generative AI dominates the headlines, the data tells a different story. According to research from Andrew Ng and McKinsey, a staggering 80% of successful AI projects in business come from supervised learning.

Why? Because it's practical, powerful, and has a clear path to ROI.

Instead of asking "How should we adopt GenAI?", the more immediate, value-driven question is: "What critical business question, if answered, would transform our bottom line?"

Supervised learning uses your own historical data to build a crystal ball for your business, answering questions like:
-Which customers are most likely to churn?
-Which sales leads are most likely to convert?
-How can we minimize inventory shrink and waste?
-What product should we recommend to this specific customer?

Solving these challenges directly translates to reduced costs, increased revenue, and higher customer lifetime value.

Getting started is more accessible than you think:
-No PhDs Required: Identify an intellectually curious employee who understands your business context and upskill them. Their domain knowledge is your greatest asset.
-Use Existing Tools: Many companies start by leveraging probabilistic features in software they already own (like Hubspot AI or Salesforce Einstein) or by building pilots in Excel or Google Sheets.

AI is a tool, and value is derived by the humans who use it. Start with a tangible business problem, secure a few small wins with clear financial impact, and build momentum. These successes will fund your long-term AI roadmap and foster a culture of data-driven innovation.

Your Call to Action for Next Week:

Don't just contemplate AI. Put it to work.
1. Identify: Ask your team: "What single insight, if we had it, would be the most valuable for increasing sales or enhancing profitability?"
2. Quantify: Estimate the dollar value of gaining that insight.
3. Act: You've just defined the business case for your first high-impact AI project.

This is how you build a practical, profitable AI strategy. It's not about chasing the latest trend; it's about solving core business problems. The opportunity to create significant value is waiting in the data you already own.

The F1 Lesson Every CEO Is Missing: Why Your Pit Crew Determines Your Customer Race

Fri, 12 Sep 2025 21:12:14 GMT

Lessons from the Racetrack

Most executives study customer experience from the safety of the boardroom.

Brad Pitt’s character in F1 did something radically different. And it gave me a new perspective on how to think about customer strategy.

The Track-Running Executive

Before every race, Brad Pitt’s character ran the entire track on foot. Not to stay fit but to understand every curve, bump, and blind spot his car would face at 200 mph.

Then something remarkable happened: his pit crew started running with him.

This is what true leadership looks like. Most executives analyze customer journey maps in meetings but have never:

Called their own support line
Tried to sign up like a new customer
Felt the frustration of a billing error firsthand

When leaders and frontline teams experience customer pain together, everything changes.

They stop debating survey scores.
They start solving real problems.
They develop a shared language around what truly matters.
They understand why seconds matter in every customer interaction.

Your Pit Crew Determines the Race

Here’s what Brad Pitt’s character understood that many CEOs miss: The driver might get the glory, but the pit crew determines whether the driver can compete.

Your pit crew is made up of:

The service rep handling a tough escalation
The onboarding specialist guiding a nervous new client
The billing team trying to fix a payment issue before a customer walks away

These moments decide the race for your business.

Yet, too often, these teams are treated as cost centers. Product, sales, and strategy teams get the spotlight while frontline employees, the ones who see every customer interaction, are left out of the room.

Here’s the hard truth: Your frontline teams know more about your customer experience gaps than anyone in your executive suite. They hear the complaints. They see the broken systems. They witness the moments when a customer chooses to stay or leave.

The Integration That Changes Everything

When leaders “run the track” with their teams and elevate frontline voices, transformation accelerates:

Shared understanding builds trust. Leaders stop seeing support tickets as interruptions and start viewing them as a window into customer reality.
Frontline teams become partners. When they have a voice in decisions, they share insights no dashboard or consultant can surface.
Customer experience becomes everyone’s job. It’s no longer a problem for one department to solve.

The Executive Challenge

When was the last time you experienced your own customer journey?

Not observed it.

Felt it.

Felt the hold times. Felt the website glitches. Felt the frustration of being passed around between teams.

And how are you including your frontline teams today? Are they simply reporting problems, or are they sitting at the table solving them with you?

Your customers don’t care about your org chart. They care about whether the person helping them in the moment makes things better or worse.

Those employees are either your competitive advantage or your greatest risk.

In F1, races are won or lost in the pit stops. The same is true for your business.

What would change if you started running the track with your pit crew?

Unlock Growth by Fixing What's Broken Inside

Wed, 03 Sep 2025 21:41:02 GMT

Your single greatest opportunity for growth isn't in a new marketing campaign. It's buried in the daily frustrations of your employees and the hurdles your customers face. Aligning the employee and customer journey is the key to unlocking it.

The concept is simple: employee friction is a direct tax on your customer experience. Every internal process hurdle, data gap, and frustrating workflow is ultimately paid for by the customer. This results in lower satisfaction, reduced loyalty, and a direct hit to your brand reputation and top-line revenue.

So, how do you systematically identify and eliminate this friction? You can leverage my AUDIT™ Framework . It begins by assessing reality on the ground and ends with a transformed culture of continuous improvement.

The AUDIT™ Framework

A ssess reality on the ground through data and stakeholder insight.
U nderstand root causes and, critically, their financial impacts.
D esign the future-state journeys that drive efficiency and value.
I mplement the prioritized, high-impact solutions.
T rack progress and transform the culture with data-driven governance.

Let's walk through the framework.

Assess & Understand: Quantify the Pain

The process begins with a pain point analysis, including cross-functional interviews. This isn't just about gathering anecdotes; it's an impactful learning exercise for leaders who may be removed from the day-to-day. As you assess pain, your employees will invariably describe both their journey and the customer's journey, offering direct insight into removing friction.

Next, you must Understand the root causes and dollar impacts behind every pain point. Is the issue rooted in training, process, or technology? Are there data gaps forcing manual workarounds? The findings must be prioritized by business impact and feasibility.

Design & Implement: Architect the Solution

Following the 'Understand' phase, we move to the strategic Design phase. This is where we architect future-state journeys and the corresponding KPIs that unite employee and customer success. The focus is laser-sharp: deliver efficiency and optimize work processes to create superior customer experiences. This could be a Customer 360 view for your frontline teams or a self-service portal for your clients. The key is to align the design with metrics that matter: NPS, CSAT, customer effort, and employee satisfaction.

The Implementation phase is where the rubber hits the road. Leaders, now armed with deep insight, can build a strategic program by funding prioritized projects and assigning clear ownership. When employees see new processes, systems, and tools come to life, they know you've listened and have their back. Remember, your employees and customers are your two most valuable assets.

Track & Transform: Drive Continuous Improvement

Finally, Track & Transform . This phase embeds data and analytics, like real-time dashboards, and a governance cadence so improvements compound over time. When the entire organization is aligned to the critical metrics that serve customers, you cultivate "act like an owner" behaviors because everyone understands their essential role in the journey.

Case Study: The $1.3M Coffee Shop

Let's apply this framework to a hypothetical coffee shop with drive-thru, mobile, and in-café channels.

Average Revenue: $1.3M / year
Average Ticket: $6.50
Pain Points Discovered (Assessment): In-house customers wait the longest, high-complexity drinks slow the drive-thru and cause errors, employees are too stressed for customer interaction, and at least 10 mobile orders are abandoned daily.

Understanding the Financial Impact: During the Understand phase, we move beyond surface-level metrics to pinpoint how operational friction impacts financial performance. With mobile orders, for instance, the transactional profit is secure. However, the real financial story is in the opportunity cost and customer risk these abandoned orders reveal.

First, each abandoned order represents wasted production capacity. During a busy morning rush, a barista's time is both the store's most valuable and most limited asset. Every minute spent making a "ghost" drink is a minute not spent serving a waiting customer in the café or drive-thru. This directly reduces the store's throughput and caps its potential revenue. We calculated this seemingly small issue costs the store nearly $5,000 annually in lost potential sales from just two peak hours a day.

Second, and more strategically, an abandoned order is a red flag for a poor customer experience that puts future revenue at risk. A customer who leaves without their order is unlikely to return soon, jeopardizing their entire lifetime value. This transforms a simple process inefficiency into a significant threat to long-term growth. By quantifying both the immediate opportunity cost and the long-term customer risk, we can build a powerful and undeniable case for change.

Designing and Implementing the Solution: In the Design phase, the store layout, staffing model, and technology are re-architected. A dedicated team and workflow are created for mobile/drive-thru orders, while the in-store staff is optimized for the café experience. In the Implementation phase, the store is redesigned, algorithms are updated, and team members are trained on new, role-specific tasks.

Tracking the Transformation and ROI: The goal of the new design isn't just "higher CSAT." It's about measurable business results. Through new dashboards, the team can Track progress:

A projected 5% reduction in drink errors saves thousands in ingredient waste.
An increased drive-thru throughput of 15% during peak hours could capture dozens of additional transactions daily, translating to significant new annual revenue.
Higher employee sentiment reduces costly turnover.

This is how the framework aligns employee satisfaction with customer satisfaction to deliver tangible financial outcomes.

Conclusion

Aligning the employee and customer journey is not a 'soft' initiative; it is a core business strategy that directly drives operational efficiency, customer loyalty, and bottom-line growth.

What is the single biggest point of friction your employees face, and what do you think it's truly costing your customers?

Ready to unlock the hidden revenue in your organization? The framework above can be adapted to any industry or operating model. Start with assessment, the insights alone will surprise you.

The Power of Customer Effort Score

Wed, 28 May 2025 20:47:09 GMT

Reducing Friction and Driving Loyalty

Last week, I had to call my health insurance company with some questions, a task I always dread. How many buttons will I have to press before reaching a live human? Once I do, will they be able to help me? Before making the call, I had already searched the company's online portal, intending to self-serve, but I couldn't find the answer I needed.

Fortunately, when I finally reached a customer service representative, they were knowledgeable and resolved my issue efficiently. Relieved, I ended the call with a simple "thank you" and confirmed that my issue had been resolved.

A few days later, I received a follow-up call asking about my experience. The questions covered politeness and issue resolution, but one stood out: "Did your service journey start with the phone call or the website?" They also asked, "Did the company make it easy for you to solve your issue?" This question struck me because it measured my effort—not just my satisfaction. It reinforced an important lesson: customer friction is a key determinant of loyalty, and businesses need to measure and address it systematically.

Why CES Matters

In the book The Effortless Experience, the authors highlight a counterintuitive truth: reducing customer effort has a greater impact on loyalty than exceeding expectations. Companies that focus on "delighting" customers often end up raising service expectations unsustainably, leading to more dissatisfaction when those expectations aren't met. The book also emphasizes the importance of refining the Customer Effort Score (CES) as a measurement tool to ensure customers understand what they are being asked and can effectively rate their experiences.

So, how did I respond to the health insurance company’s survey? I gave a 5 out of 7 . Even though my issue was ultimately resolved, I had to navigate a website without finding answers before making the call, an extra step that increased my perceived effort. And that’s the key: perception matters. Customers often believe they’ve exerted more effort than they actually have, which is why businesses must proactively reduce both actual and perceived effort.

How to Reduce Customer Effort

Reducing customer effort leads to better business outcomes. Research from Gartner shows that lowering effort increases repurchase rates, lowers service costs, and improves employee retention. Here are some ways to improve customer experience by minimizing effort:

Reduce Journey Steps – Eliminate unnecessary steps in customer interactions to streamline processes and remove friction.
Eliminate Repetitive Feedback – Customers shouldn’t have to repeat information they’ve already provided. Implement chatbots, autofill features, and single-click interactions to make experiences more seamless.
Pre-fill Customer Information – Use existing data to auto-fill forms and reduce the need for repetitive data entry.
Anticipate Customer Needs – Offer clear guidance, instructions, and next steps before customers even ask.
Show Progress Indicators – Let customers know where they are in a process, so they feel in control and informed.
Personalize Recommendations – Leverage customer data to quickly guide them to the most relevant information.
Optimize Digital Interfaces – Ensure websites and apps work flawlessly across all devices and load quickly—speed significantly impacts perceived effort.
Minimize Decision Fatigue – Simplify complex processes by curating options and reducing overwhelming choices.
Deliver Empathetic Support – Train agents to recognize and alleviate frustration. Even when delivering bad news, an agent can reduce effort by providing clear alternatives and resolutions.

How CES is Calculated

The authors of The Effortless Experience recommend measuring CES using a 1-7 scale, where 1 = very difficult and 7 = very easy. The score is then calculated as follows:

Total customers who rated the experience as easy (scores of 5, 6, or 7)

Total survey respondents

This results in a 0 to 100 score, with higher scores indicating lower customer effort.

Another approach is a Likert scale, where customers select from responses like "Very Difficult" to "Very Easy." Emoticons or images can also be used for surveys conducted via email.

Where CES Can Be Applied

CES can be used across industries and touchpoints to improve customer experience. Here are some examples:

Service Industry – Measuring the effort required to resolve billing issues, schedule appointments, or receive support. For instance, a doctor’s office could use CES to assess the difficulty of booking an appointment or understanding a bill.
Technology – Evaluating how easy it is to use software or online platforms. A software company could measure CES to gauge the effort needed to generate a report or download data.
SaaS/PaaS – Assessing onboarding, feature adoption, and customer service interactions. A SaaS company might use CES to track how easily customers troubleshoot technical issues or navigate self-service portals.
Subscription Services – Understanding how easily customers manage subscriptions, billing, or usage. A cloud storage provider could measure CES to evaluate the ease of tracking usage or adjusting storage plans.

CES and Customer Channel Switching

One of the biggest indicators of high customer effort is channel switching, when a customer moves from an online portal to a phone call because their issue wasn't resolved. When customers must escalate their issue across multiple channels, their perceived effort skyrockets. Companies must design seamless experiences that anticipate customer needs and deliver the right support at the right time, in the right channel, minimizing the need for extra steps.

Why You Should Start Measuring CES

If your company isn’t measuring Customer Effort Score (CES) yet, you should be. Understanding the friction customers face helps eliminate pain points, making them more likely to stay loyal and spend more money with your business.

Curious about how to implement CES in your organization? Let’s talk!

Beyond the Buzzword: What it Really Means to be Data-Driven

Wed, 28 May 2025 20:47:07 GMT

This is a subtitle for your new post

The Current Reality

Many organizations claim to be “data-driven,” yet few deliver on that promise. In practice, data is often used to validate existing beliefs or simply report on what’s already happened. The explosion in data collection has far outpaced our ability to extract and act on meaningful insights. This disconnect creates the illusion of a data-driven culture, one that measures everything but changes nothing.

What True Data-Driven Decision-Making Looks Like

In a genuinely data-driven organization, decisions are grounded in insight, not instinct. Data shapes the strategic direction forward, not just day-to-day execution. Employees at every level within the organization can access the data they need when they need it! Data is leveraged to improve customer experiences and business outcomes.

If it's not measured, it doesn't get done.

Core Elements of a Data-Driven Culture

To move beyond lip service, organizations must embed the following elements:

Data democratization, paired with the right level of data literacy across roles.
Action-oriented analytics that drive real business decisions, not just dashboards or reports.
A culture of experimentation, where hypotheses are tested, and learning is valued.
Feedback loops to assess the impact of data-informed decisions and adapt accordingly.
Customer-centric metrics that connect directly to business value and outcomes.

Common Pitfalls to Avoid

Even well-intentioned efforts can go sideways. Be on the lookout for vanity metrics that look nice but have zero relevance to the business. Analysis paralysis can occur if there is so much data that it overwhelms an individual's ability to make a decision. Siloed systems can be problematic and block cross-functional insights. Inaction may occur when data insights challenge the status quo or internal politics.

How to Make the Shift

Becoming truly data-driven requires intention, not just infrastructure. Here’s where to start:

Lead with the right questions, not just the data on hand. Most business needs can be address with a clear, core set of questions.
Build decision frameworks that clarify which data matters most and why.
Invest in data literacy, tailoring programs to the needs of different functions. This will ensure everyone interprets the data through a common lens.
Establish enabling governance, ensuring data is secure, trusted, and accessible without unnecessary friction.

Ready to Turn Insights into Impact?

I help organizations bridge the gap between data potential and business outcomes. Whether you're just getting started or need to accelerate progress, I partner with teams to build the strategy, structure, and culture needed for real data-driven transformation.

Why Culture Starts at the Top and Shapes Every Customer Experience

Wed, 28 May 2025 20:42:34 GMT

Culture is Cultivated

I’ve been thinking a lot about culture lately.

With headlines full of companies flattening their organizations and expanding spans and layers, I can’t help but wonder what will this mean for culture? What does it mean for the people at every level who show up each day to do the work?

I’ve also been reflecting on a company I worked for that had a culture I was genuinely proud to be part of. It was a private label health and beauty company called Vi-Jon. Midway through my career; I joined their team. On my first day, I was sitting alone at lunch when someone sat down across from me. I asked what he did at the company. “I’m the CEO,” he said. His name was Jerry.

That simple act, sharing lunch with someone new, spoke volumes. Jerry was like that. He walked the floors of the plant. He knew people’s names. He had real conversations. He was sharp, kind, and steady. You wanted to do your best work around him.

The team at Vi-Jon was scrappy. We figured things out. We made things happen. And we showed up for one another. I’ll never forget the time I had to get a Walmart bid package out the door but was short-staffed. Two of the company’s VPs, one from Sales Ops and one from R&D, stayed late with me, hand-filling bottles of soap and other samples. That’s what culture looks like in action.

I still think about those days. I miss the camaraderie. The clarity. The energy that came from working alongside people who believed in what we were doing.

The word “culture” gets thrown around a lot. But what does it actually mean?

Culture isn’t a slogan on the wall or a PowerPoint slide. It’s how people behave when no one is watching. It is the shared beliefs, values, and unspoken norms that shape every interaction from internal meetings to customer conversations.

At its core, culture is built on underlying assumptions, those deeply held beliefs about how things really work in an organization. These are often invisible, but they’re incredibly powerful. And they’re hard to change. They’re the heartbeat of the company.

So how do you build a culture that drives performance and creates meaningful customer experiences?

1. Leaders set the tone. Always. Leadership defines culture. Leaders are the blueprint. People don’t follow words, they follow behavior. If leadership says one thing but does another, employees notice. Culture crumbles. But when leaders show up consistently, with integrity and humility, trust follows. Culture follows.

2. Managers are the bridge. Senior leaders may define the vision, but managers make it real. They’re the ones employees look to every day. If middle managers aren’t aligned or worse, if they’re checked out, the culture breaks down. That is why training, coaching, and supporting managers isn’t optional. It is essential. Managers must embody the values, give honest feedback, and keep the lines of communication open.

3. Internal culture shapes external experience. How employees treat each other shows up in how they treat your customers. Internal alignment and trust lead to better decisions, stronger collaboration, and more innovation. Companies with strong cultures don’t just feel better to work in, they deliver better business outcomes. On the flip side, a toxic culture like “growth at all costs” or internal competition creates dysfunction that seeps into every customer touchpoint.

Look at Wegmans. Their culture is rooted in caring: caring for employees, for customers, and for the communities they serve. It is no accident that they have built a fiercely loyal customer base and remain one of the most respected brands in grocery retail. When people feel valued and supported at work, it shows up in every aisle, every interaction and every decision.

Culture is not a program or an initiative. It is built every day by what leaders do and what they tolerate.