Thumbs down

Why do B2B customer experience teams fail?


Why do so many customer experience teams fail to produce results in a business-to-business environment? Maurice FitzGerald pulls on his 40 years of work in the domain to list the five most common reasons. 

6th Jan 2023

A pessimistic Forrester report and a recent post on LinkedIn made me research and reflect on the main reasons customer experience teams’ work fails to produce results in a B2B environment.

Such failures often lead to the elimination of the CX team, sometimes disguised by phrases such as “We have decided that CX is everyone's job, and have therefore eliminated the central CX team, and the CX teams in each business unit and function.” Almost makes the decision sound rational, doesn't it?!

I have decided to document the top reasons for such disasters based on research and on my own 40 years of experience in the domain. And yes, I have even witnessed my own CX work being eliminated, so believe my insights may be helpful. 

1. No sponsorship from the top

I am defining sponsorship as having the CEO or a member of the executive team who provides you the funding, project managers, and other resources that you need to drive strategic CX improvement initiatives across your company. If you don't have such sponsorship, you will fail.

If you currently have a job as a CX leader, you may be incorrect in thinking that this means you have sponsorship. For example, perhaps you report to marketing and the general consensus is that every marketing team should have a CX person to do customer surveys. That is not the same as real sponsorship. A simple test would be to ask for three project managers, a comms person, and $400k to implement a major improvement project and see what happens.

Note that I am talking about getting truly deep sponsorship in the first place, rather than preserving it once you have it. In commercial organisations, the people at the top are primarily interested in money. The top investments they believe will significantly improve growth, profits, and the company share price are good and will get the necessary resources. Everything else is nice to have, at best. Most CX teams and programmes fall into this ‘nice to have’ category. The team name may prevent its total elimination, but the relationship between the work the team does and financial outcomes has not been demonstrated. And they have never proposed a major investment in the past. 

Can surveys help?

Proof of the value and impact of CX is key, and the key to proof lies in the data. The traditional (failed) CX approach is to rely on survey data. But they are seen as interesting market research and a source of internal scorecards, not as the reference for the next major corporate strategic initiative. Surveys are interesting, but are not good enough to create the level of sponsorship we need.

There are three sub-problems here:

  1. Surveys only cover a fraction of the customers and even within that fraction, the correct individuals may not be be responding. For B2B, the respondent profile is critical to survey accuracy.
  2. B2B relationship survey results tend to only be updated annually, quarterly at best. This means they happen outside the financial reporting cycles.
  3. Surveys are disconnected from the operational realities, meaning disconnected from the hundreds of operational performance metrics that your various teams use to measure themselves every day. Executives and team leaders simply consider the data to be of poor quality. They consider other data sources to be far more reliable.

So, how exactly can these three problems be resolved? 

Get sponsorship by improving the data sources and proving financial linkages

The key to success is to base CX work and corporate-wide improvement priorities on the data sources that everyone already trusts: operational data that comes from the various systems your company uses to manage its business. Match the hundreds or thousands of operational data points and trends you have against the outcomes for the same customers. Which data points and trends seem to guarantee contract renewals, and which do not. Which mean customers buy more, and which operational experiences push customers away. You can cover 100% of your customers using modern AI software such as OCX Cognition's Spectrum AI. 

Once you have this data it is extremely obvious which parts of the company and which teams are performing well and which perform poorly. Now you know the basics of how to get compelling data that will ensure sponsorship for your work from the people at the top. Money matters, and it is indeed all about the data.

2. No data

Let’s suppose you have sponsorship from the top, simply based on what you have promised is possible. That confidence in your promises cannot endure without proof. The proof requires data, but what data, precisely?

What about survey data?

You may have survey data, but for the vast majority of B2B companies on our planet, that has made no difference. Yes, you have been able to close the loop with some unhappy customers. However, you have not been able to use it to secure investments in strategic CX improvement initiatives. In such situations, I hope you will recognise that the strategic result is the same as having no data at all.

Here are the main problems with relationship survey data (Don't even get me started on the problems with transactional survey data!):

  • Low frequency: Results are available annually or quarterly, while every other management report is available daily, weekly, or monthly at worst.
  • Low response rates: We consider 20% response rates to be ideal. How can we possibly consider ourselves credible with no information whatsoever about 80% of our customers?! That's simply not good enough.
  • Massive sample bias: We consider all respondents as equal. The 20 companies that provide half our revenue get exactly the same weight as another 20 that constitute 0.001% of the total. There are other biases, but this it the biggest one.
  • Wrong respondent: Particularly when the sales lead is the person who decides who should receive a B2B survey request, only favourable people tend to be selected. Let's take outsourcing services as an example. The real decision-makers are probably the CFO and the procurement manager. We have often found them to be completely missing from respondent lists.
  • Incomplete lifecycle coverage: This overlaps with the point about respondents. The people who respond to relationship surveys will provide information only about the part of the whole customer journey that they have experienced. What about the rest?

Your current data is incomplete, discontinuous, and disconnected

The result of all of this is that your current data is sadly lacking. You are missing essential information. The data you have is not available on a continuous basis, and the format is disconnected from the operational language, metrics, and realities that your company's teams face every day. That's bad. It has the same strategic effect as having no data at all!

The solution is to use your existing operational data from the systems used to manage customers across their entire journey with your company.

[Click to enlarge]

customer journey


This diagram is for a typical SaaS customer journey and I am sure it is meaningful for other industries too. Each team that manages a stage in the customer journey uses its own IT systems. Yes, Salesforce and a couple of others may be used for more than one stage of the journey. Let's me try to express this in the way it first struck me when our own Brian Curry made me look at it this way: nobody, absolutely nobody, looks at the entire set of data across the customer journey to try to establish what works and what does not. Which of these data points and trends are most likely to give you a ‘customer for life’ and which are most likely to drive them away from you, into the arms of a competitor.

Nobody has looked at it this way because nobody cares about it this way, below the CEO. As a CX leader, you need to make yourself the exception to this rule. You need to implement the necessary software to gain meaningful end-to-end insights. It's doable within a single quarter. As McKinsey put it: "Why use a survey to ask customers about their experiences when data about customer interactions can be used to predict satisfaction?"

The ‘no data’ problem is real. And by data, I mean complete, continuous, and connected data. Data that covers all customers, is constantly up-to-date, and is directly connected to the way each team in your company measures itself. If you can't become complete, continuous, and connected, I believe you are guaranteed to fail. So, why not get started today!

3. No strategy

Most B2B CX programmes simply lack a strategy. Please don't be tempted to think that a strategy and a plan are the same thing. Strategy and planning are almost complete opposites, while remaining related to each other. The reason is simple: strategy is a divergent process, while planning is convergent. Strategy work identifies new things to do and is therefore about things that you do not currently do. Planning is about putting things in place that have already been defined. Strategy is a creative process and planning absolutely, positively is not! 

So what is strategy?

Strategy is about how to deploy your scarce resources to win. You don't have unlimited resources and can't possibly do everything you would like to do. You have to be selective. There are two steps involved in creating a good strategy:

  1. Situation analysis: understand what is going on and what will change.
  2. Make your strategic choices.

Generally speaking, the thing that most people get wrong when creating a strategy is the situation analysis, and that is what I want to cover in a bit more detail. Implementing the strategy is an entirely separate discussion.

Situation analysis

A great situation analysis has six components. I have written extensively about the subject elsewhere and have nothing new to say about four of them. However, I have quite a lot to say about customer and partner insights.


In the old days (meaning up until about two years ago) I used to say that the best way of performing a full situation analysis on your customers and partners was simply to ask them, meaning to survey them. I don't believe that is sufficient anymore, simply because it is difficult to get an accurate and complete picture from surveys, even in B2B, where response rates tend to be higher. If you don't believe this, ask yourself what you know in detail about one of your customers who has not responded to your survey.

So, if surveys don't provide enough insights, what does?

That's pretty easy: your operational data does. However, an average B2B company tends to have about 2,000 data points for each customer and partner, covering the entire customer journey. That data is an unbiased source of potential insights. Customers who are loyal and buy more tend to have quite different operational data trends from those who buy less and less, and ultimately go to a competitor.

In our experience implementing our own Spectrum AI software to analyse such data, we have been surprised by the number of surprises we get. Yes, it is true that customers' intuitions about problems and success factors are sometimes entirely correct. However, they are usually mistaken, or at least incomplete. For example, it is common for customers to believe that the most important insights will come from support system data, but this is rarely the case. Finding major improvement opportunities in Salesforce, Gainsight, Jira, or SAP data is quite common. And naturally, finding them when you have plenty of time to address them improves retention and margins.

Best of all, finding new operational issues that are common to many customers allows you to move away from reactive fire-fighting mode, into proactive problem solving.

Strategies are about choices, and data is central to making the correct ones. Surveys used to be the main reliable tool to get input about what B2B customers and partners want from us. Operational data is an improved source of insights, not least because it can provide us with reliable information about customers who do not respond to surveys.

4. No plan

Suppose you have decided to implement three major strategic initiatives to improve customer experience. You may think you have a plan, but you probably don't, or at least not a good one. I learned some of what follows quite painfully. I had formally been managing projects for something like ten years when my manager asked to see me in his office in a suburb of Munich. We talked about the main project I had to lead at the time. About ten minutes into the conversation, he said “Maurice, you seem to know almost nothing about project management.” I was upset, to say the least. In any case, he explained why he thought that, and I have to say that he was right. What follows are some of the things I was missing.

You need formal training to understand even the basics

Just like most engineering and business school graduates, I had never had formal project management training of any kind. To me, the worst problem I had was that I did not understand the difference between a project task and a deliverable. 

A task is the work that needs to be done. A deliverable is how you can observe completion. Tasks start with a verb. Deliverables start with a noun. A task might be to create a new report. The deliverable is the first report. The plan should also name the single person who has to approve the task as complete. In many cases, data is essential to be able to observe a deliverable, for example if the task is to improve a specific operational deliverable you need reliable data on that deliverable, not just somebody's opinion.

If it takes too long to produce the first major deliverable, your strategic initiative is certain to be cancelled.

The task might be to create a new order picking process in a warehouse for the 100 most popular products so they can be shipped during the afternoon if the order is received in the morning. You need reliable data to prove that the deliverable has happened and that the same-day shipping is now in place.

Also, if you do not directly involve the people who have to implement your plan in its creation, failure is guaranteed. They will always feel somewhat offended that someone they perceive as knowing nothing about what their team does has created work for them, and they will simply reject the project plan you have given them to implement.

Dependencies on other teams have to be discussed face-to-face

In this day and age, it is common that even major projects are kicked off online. If a new initiative involves seven departments, a PM will usually attend from each department. After an initial discussion, the overall PM then closes the first meeting and holds separate calls with each of the seven PMs. But this process creates a huge issue. Each team will almost certainly have listed dependencies on deliverables from other teams in the project plan. But the other teams will know nothing about those deliverables.

The only way I have found to resolve this is to hold kickoff meetings face-to-face. I feel that trust is something that can only be established by looking at another person directly in the eyes. So, don't hold virtual kickoffs - do them in person. Add what I call a ‘speed dating’ session to the agenda. Have the PMs rotate around the room so that each and every one can discuss and agree the dependencies with the other PMs. As always, it is critical to agree on the data that will be used to define the dependencies as complete.

While I could go on at length, I just want to provide one more piece of input. If it takes too long to produce the first major deliverable, your strategic initiative is certain to be cancelled. The reason is simple: senior leaders are under pressure to demonstrate progress each quarter, at the very least, and so must you, with the data that proves progress has been made. And of course, if you take too long to deliver, there is certain to be some sort of leadership change and you will lose sponsorship.

I have provided some indicators that should help you to understand whether you have a plan or not. The question is not a simple one, and you will need to make it all as clear as possible to your senior leaders. Basing your completion criteria and success criteria on observable operational data trends and numbers is critical to success. In the process I have been describing so far, we have defined the customer experience improvements that are required in terms of concrete operational numbers, rather than in relatively abstract NPS or CSAT terms. These operational numbers are something that your leaders, and indeed your own project team, can observe every day. Everyone, including you, will know you have a plan, and will be able to observe progress easily.

Nobody cares

You are working hard but each time you present to senior leaders it seems like you are starting from scratch. You have to explain your survey method and NPS results. It's even worse if you use a metric other than NPS that nobody understands without a new explanation every time you present. Let's also suppose everything I covered above is in place, but you still can't deliver results. The most likely explanation for all of this is that nobody cares... And the fundamental reason nobody cares is usually that your efforts at effective communication have failed.

No emotion

Regular readers know I am a big fan of Daniel Kahnemann's behavioural economics writings, and his ‘System 1’ and ‘System 2’ thinking. System 1 is our intuitive, emotional reaction to information we receive. System 2 is all about rational thought. The key point is that once our System 1 has jumped to a conclusion, System 2 does not engage at all, meaning we are totally unreceptive to rational explanations. What this means is that you need to lead with a System 1 / emotional message, then back it up with data. If you lead with the data, your audience will jump to its own conclusions while you are presenting and you will not be able to change their minds.

The fundamental reason nobody cares is usually that your efforts at effective communication have failed.

Customer quotes that support the argument or investment request you are about to make are a great way of accomplishing this. Such quotes remain in people's heads far longer than your numbers.

Another great way of getting people to want to help is to ask for their advice, and I mean that you should do this explicitly. “I have quite a lot of compelling data here today, but that is not my main reason for being here today. I really need your advice.” 

Wrong type of metrics and data

For many years, I made annual presentations about CX results to various HP business units and other teams. With perfect hindsight, I realise that a major mistake I made was to speak in my language, using my metrics, rather than using their data and metrics. I used terminology the various teams only heard rarely and on which they were not measured. At the time, I didn't really have a choice, as I only had NPS and associated driver data, and had nothing at all that used their everyday operational metrics.

Again, we need to think of communication as something that happens at the receiving end. If you are presenting to a sales or business management audience, forget about NPS categories. Tell them what you have found out about ‘Q3 at-risk revenue’ segmented by account type. Since you have the necessary information about every single customer, you can answer specific questions too.

Summary of the five reasons

Here is the way I look at the five main reasons for B2B CX failure that I have covered in the last couple of weeks. It should be clear that I consider data to be at the center of everything. Incomplete data, data that is not available continuously, data that is disconnected from the way team operate on a day-to-day basis. These defects in data are usually the main reason you have no sponsorship, have not created a real strategy, don't have a concrete implementation plan, and find yourself in the situation where nobody seems to care.



Seems pretty depressing, right? It does not have to be. In my next article, I will attempt to set out how to fix these problems quickly in 2023.

This article adapted from a series that Maurice published on LinkedIn


Replies (0)

Please login or register to join the discussion.

There are currently no replies, be the first to post a reply.