These Are Your Customers’ Service Preferences and Frustrations

Like pearls develop in oysters as protection against irritants, so customer frustrations can develop into pearls in the form of insights that help to improve your service setup.

Here at Userlike, we’re interested in understanding how customers think about customer service. What are their biggest frustrations? And what communication channels do they prefer in different situations?

We set up a study to find answers to these questions, hoping this could help us to further develop our own communication solution . But we soon realized the answers are interesting for anyone aiming to improve their service setup. It’s for those that we’d like to share the results of our research in this post today.

Take a look at what we found out.

The average customer service can be strongly improved

Perhaps the most surprising result we got is that a whopping 87.6% agreed or strongly agreed that customer service departments could be strongly improved. It’s a bold statement that they could be strongly improved, which is more than just idealistic thinking (“There’s always room for improvement”). In comparison, only 5% disagreed or strongly disagreed with that statement.

pie chart showing percentages who thought customer support departments could be improved

There were more diverse opinions about the quality of customer service. 30.5% agreed or strongly agreed that the customer support department of most companies are of good quality. Meanwhile, 38.1% disagreed or strongly disagreed and 31.4% felt neutral about it.

pie chart showing percentages who though customer support departments are of good quality

Customers already online prefer to use live chat

The bulk of our survey was made up of situational questions. These were questions in which we asked the respondent to picture a specific situation and then asked her to choose what contact channel she would use to get in touch.

The results showed that the majority of customers would use live chat when they’re already online. 66.7% of respondents answered that they would choose to use live chat when they’re shopping online to ask about a product. Email and phone both earned 14.3%, making live chat here the clear winner.

pie chart which channel to use when shopping online

The next situation we presented was, “You are on a website, looking to get a new insurance. You need advice on which plan to choose. Which channel would you use to contact support?” Live chat earned 55.2% of responses, while phone trailed behind with 28.6% and email, with 11.4%.

pie chart which channel to use when looking online for insurance

The trend continued when we asked Mturkers which customer service channel they’d use when they’re on a website wanting to book a hotel room for an upcoming vacation. Check out the results:

pie chart which channel to use when booking a hotel online

In all 3 scenarios, respondents seem the least interested in using mobile messaging or social media to contact support, each gaining about 5% or less in every situation.

Looking for better customer relationships?

Test Userlike for free and chat with your customers on your website, Facebook Messenger, and Telegram.

Read more

Live chat also wins offline

We also presented the Mturkers with a situation in which they weren’t already online and had a question for customer service. Surprisingly, the majority (49.5%, to be exact) still voted live chat. We expected mobile messaging to rack in the most votes, but it actually only got 5.7%.

pie chart which channel to use when checking your order status

My hypothesis for this outcome is twofold. Firstly, perhaps we didn't phrase the question to be clear enough that it was an “offline situation.” We didn't indicate that the situation was online, but didn't exclude it either.

Secondly, the problem with mobile messaging might be that there's no clear connection point. If you don't already have a contact, you have to find the company’s WhatsApp number or Facebook page first before messaging them, which requires time and a bit of searching.

More likely to complain via email

We were also interested in which channels customers would use to complain. Here are the results:

pie chart which channel to use to complain about quality

It’s interesting to see that the majority of respondents would choose email. Perhaps shyness is a factor and email is a good way of circumventing speaking directly with a person, while still giving you the courage and space to speak your mind.

We do notice a jump in use towards social media, which isn't shocking, considering social media is a beloved tactic to get attention from companies. In fact, another research showed that 45% of consumers said they would call out a company via social media because of a bad product experience .

Moderate feelings about email

We focused a couple of questions on email, based on typical problems we know to be associated with this service channel. The results weren’t especially polarizing, as expected. Here are the results we got:

pie chart about long response wait times for email
pie chart effort of findng a companys email address

Phone troubles

Responses to the situational customer service questions showed low interest in calling customer support. Our questions directed towards the frequency of frustration gave us some clearer insight as to why.

41% of respondents said it’s frequently a problem that they are put on hold for a long time when calling customer service and 16.2% said that it’s always a problem. For some comparison, only 1.9% said it’s never a problem.

pie chart waiting on hold for too long

Customers are repeating themselves too often

Repeating questions or issues when calling support also seems to be a problematic area. No one said that it’s never a problem, while 29.5% said it’s frequently a problem. It also showed in the responses to the short answer.

pie chart repeating question on phone too often
quote from short answer question

Live chat finds an easy solution to this problem. When a customer is chatting with an Operator and is then forwarded to another Operator, the second Operator automatically receives a transcript of the chat conversation. That way, live chat gets rid of the need for a customer to repeat herself. Simple as can be.

Respect in customer service is lacking

Friendliness and respect is a crucial aspect in customer service , so we wanted to find out how customer’s view this issue. We presented the statement to respondents, “When I reach out to customer support, I do not feel respected by the customer support agent.”

We’re happy to see that 32.4% voted that it is never a problem. However, 51% voted that it’s sometimes a problem, 10.5% voted that it’s frequently a problem, and 5.7% voted that it’s always a problem. While it doesn’t appear that disrespect in customer service is an overwhelming problem, it does look like it could use some attention.

pie chart respect by agent

Short answer insights

The next part of our research was a short response question: “If you could give a tip to the average customer support department out there, what would it be?” A pretty brief question, but we were pleased to see that the Mturkers really took their answers into careful consideration.

For the ease of understanding (and not copying the 105 responses we got into this post), I separated the responses into categories where I saw general themes collecting. Below, I’ve listed these categories and shared a couple of responses I felt were most reflective of the group.

Employee training

Make sure to train your employees about the problem and if they don't know the solution, then please ask from a senior executive and answer it right away, don't forward or redirect my call.

Communicate with each other better. Each representative should be on the same page as the other. Make notes on the file.

Answer questions as quickly as possible and have agents fully trained to handle any and all issues that may arise for customers.

Humaneness and empathy

Listen to your customer and divert from the script once in a while when needed.

Do not reply with generic, pre-written answers. Tailor your response to each situation and customer. They will be a lot happier with your service.

Put yourself in the shoes of the person making the complaint. Make an effort to describe your thought process as you go along.

Do your best to be kind and courteous even with difficult customers.

Waiting times and speed

Putting people on hold for a long time shouldn't happen in this day and age. It’s the main reason I hate calling customer service and rather just chat online. I understand that companies may be understaffed or just busy, but I don't think it’s acceptable to make someone wait over an hour just to get ahold of someone. Also companies need to learn to pass on information to different departments correctly. I can't count how many times I spoke to one person who fixed the problem, then later on receive a phone call from another rep who thinks the problem still isn't fixed.

Make your customer service more accessible (shorter wait times) and more quick to respond to issues and inquiries.


Like our last research , we went through Amazon’s Mechanical Turk to send out our survey that we created using Google Forms . In total, we collected 105 responses (the odd 5 were due to an over calculation) and paid about 300 euros.

We used different styles of questions to get a well-rounded collection of responses. The first kind were “quality control questions.” Whenever you’re using a survey to do research, it’s a good idea to have quality control questions to notify you of any responses that might be compromised.

These questions are meant to be similar yet contradictory so that you can easily tell if the respondent wasn’t reading closely enough or just wanted to rush through to the end. It was difficult using contradictory questions for this survey because all of the questions we asked were opinion-based and not linked to each other. Meaning that the answer to one question didn't necessarily depend on the answer of another. However, we used two questions to help notify us of any surveys that might not have been carefully considered:

  • “The customer support department of most companies could be strongly improved.”
  • “The customer support departments of most companies are of good quality.”

We found it acceptable to agree with both, so, we based it on extremes. If one person strongly agreed that they were of good quality and also strongly agreed that they could be strongly improved, we could tell that they probably shot through the survey carelessly— and thus rejected the survey.

The second type were situational questions, which I’ve briefly outlined earlier. We asked Mturkers to imagine different scenarios and to tell us which customer service channels they would use.

Thirdly, we used frustration frequency questions. We asked how often typical frustrations occur, in general, when contacting customer support. The questions were phrased to be clear, basic, and relatable to anyone.

Lastly, we included a short response question, “If you could give a tip to the average customer support department out there, what would it be?” Unlike the rest of the questions, which were multiple choice, this one was free-response. The respondents could write in whatever they wanted, enabling them to give us unhindered, personalized answers. Overall, this was the most effective method to ensure quality control for the responses we collected.


We didn't target a specific audience of respondents, but asked them in the survey to select their age range and gender. Considering that our survey wasn’t really tech-related, we expected to get a well-rounded range of respondents.

In reality, about 75% of the respondents were male and almost 75% were 37 years old or younger. I questioned the skew towards young males and did some research. A study done by the University of California, Irvine shows that 62% of Mechanical Turkers are under 30, which explains a larger group of young respondents. They also found, however, that 55% of Turkers are female, which doesn't explain why we received a much greater male response.

pie graph of the gender percentages
pie graph of the age percentages

Limitations & further research

As with any research, we faced limitations that might have influenced responses in one way or another. The first is the recency effect , which is the phenomenon that you’re more likely to remember events that you’ve experienced recently.

So, if a respondent had a seriously negative or positive experience with customer service just a few days prior to taking our survey, this might have influenced her or his overall perception.

I also anticipate that cultural differences could be a limitation here. In the same study from the University of California, Irvine, researchers learned that 57% of Turkers are from the U.S., 32% from India, 3% from Canada and the other 8% coming from various countries like the U.K., Romania, and the Philippines.

In a previous post , we wrote about how culture influences what we expect from customer service and how cultures provide customer service differently. Consequently, this probably influenced how they answered our survey.

We’re still curious about how much service perception differs by age, gender and culture. Also, considering how greatly speed and wait time influenced perceptions, it would be interesting to find out how long people are actually willing to wait until they’re helped and how much that differs per channel.

Is there anything you’re still curious about or any research you’d like to see us do? Reach out to us via Twitter and send us your suggestions. We’re looking forward to hearing from you!