Customer perceptions of personalised digital prompts

Question 1

1. Key Messages

Accepted Answer

HMRC’s Digital Prompts Programme aims to reduce tax errors made by customers submitting tax returns by presenting prompts to taxpayers as they make submissions. This research aimed to understand whether there are advantages to using prompts that are personalised to the customer, and whether these outweigh the costs and potential risks.

Customers interviewed felt that basic prompts, which contained no personalisation, were too generic to be useful. Positive behavioural responses to prompts, such as double-checking figures or completing returns more accurately, was thought to be unlikely. Customers viewed basic prompts as an attempt to catch them out, rather than to help.

However, customers felt that transaction prompts, which drew on third party data to present them with precise amounts they had earned, indicated an assumption of deliberate wrongdoing and were unfairly accusatory. The research also found that these prompts had the potential to lead to a range of negative behaviours, including deliberate underreporting and reputational damage to HMRC.

Overall, customers preferred the use of source prompts. These prompts simply informed customers that HMRC receives data from specific sources that the customer uses, such as online sales platforms, but did not specify precise amounts earned. The research indicates that these prompts do have the potential to lead to positive behaviours, including double-checking amounts and reading additional guidance.

Customers felt that prompts of all types should be framed as a genuine attempt to help rather than an accusation of wrongdoing, should be relevant to their specific circumstances rather than generic, and should be informative by providing access to new or useful information. A broad range of factors were identified that could help to achieve this:

including links to further guidance
using data from prior returns rather than third parties, where possible
appearing in response to customer actions (for example, where an incorrect figure is entered)
providing reassurance to customers when they are completing the Income Tax Self Assessment (ITSA) accurately
using Plain English
using a helpful and supportive tone, rather than indicating wrongdoing
using friendly formatting (for example, avoiding red text)

Question 2

2. Executive summary

Accepted Answer

2.1 Overview of the research and aims

HMRC is committed to using new technology to create a more effective tax system and help customers get their tax right the first time. As part of this, HMRC is undertaking a Digital Prompts Programme which involves presenting nudges to customers as they complete their tax return, highlighting potential mistakes and directing them to relevant guidance. While prompts are regularly used by HMRC, data mastery now allows use of more personalised ‘intelligent’ prompts that draw on a range of data sources. However, HMRC required a more detailed understanding of whether, and how, personalised prompts can help customers get their tax right the first time. Therefore, HMRC commissioned the National Centre for Social Research (NatCen) to carry out qualitative research into customer perceptions of personalised prompts and their reported behaviour as a result of viewing digital prompts.

2.2 Methodology

The research involved 45 in-depth interviews with customers who had completed an online ITSA in the last 12 months without the help of a paid intermediary. The interviews involved the use of hypothetical taxpayer scenarios and interview stimuli, to present participants with prompts featuring different levels of personalisation. The research used purposive sampling focused on certain customer groups of interest. These were landlords, those with unearned income (such as from a property sale), micro-businesses and gig economy workers. While the report presents the views of those interviewed, these are expected to reflect the range and diversity of views within the ITSA customer population.

2.3 Key findings

What are customers’ perceptions of the intended purpose of personalised prompts?

Customers were asked about their responses to a range of prompts, beginning with prompts with no personalisation (basic prompts) and followed by prompts that had been personalised in the following ways:

using data held by HMRC from previous tax returns (prior returns prompts)
using data from a third party specific to the individual customer and naming the data source, but not including amounts (source prompts)
using data from a third party specific to the individual customer and including specific transaction amounts (transaction prompts)

Customer perceptions varied depending on the type of prompt presented and were influenced by 3 main factors:

The source of data

Customers generally supported the use of data from prior returns to make the ITSA quicker and easier to complete. Use of third party data was also supported where customers were informed about its use beforehand, it allowed customers to double check their own records and incorrect figures could be easily corrected. However, customers had some concerns about the use of third party data. Participants expressed discomfort with the extent to which HMRC could access information about their income. Customers, particularly micro-businesses and gig economy workers, wanted to be explicitly told by third party platforms that their data would be shared with HMRC. Across all customer types, participants had concerns about this level of data sharing leading to increased data security risks.

The level of personalisation

Customers perceived personalised prompts more positively than basic prompts, which they saw as too vague in language and intention to be useful. Customers saw potential benefits to personalisation, including guarding against accidental mistakes or reducing the cognitive burden of completing the ITSA. However, there was a threshold of personalisation where customers began to feel prompts were accusatory or intrusive. Customers particularly had this response to prompts that presented multiple specific personal details together, such as the address, date of sale, and transaction value for the sale of a second home.

The tone and formatting

The tone and formatting of prompts were also important factors in how prompts were perceived by customers. Customers responded negatively where prompts indicated wrongdoing, preferring prompts framed as information to help customers complete the ITSA or fully meet their tax obligations. Black or blue text was also preferred over red, which was spontaneously interpreted as a warning or an indication of wrongdoing, regardless of the content of the prompt itself.

Despite identifying possible benefits of using personalised prompts to the customer, participants still viewed prompts as intended to increase tax revenue for HMRC and reduce tax fraud, rather than to help the customer get their taxes right.

Do personalised prompts produce more positive behavioural responses than basic prompts? And if so, how?

For HMRC, prompts have 2 key purposes:

to help customers to report accurately and avoid error
to encourage those being fraudulent to report correctly

Customers felt that basic prompts were too vague in language and intention to result in either of these intended behaviours. They felt that prompts in general would be unlikely to dissuade fraudulent customers from underreporting, but that transaction prompts were most likely to lead these customers to rethink their actions. Experienced and confident customers tended to see all prompts as irrelevant to those intending to report accurately, while less experienced or confident customers would recheck their own records or read the guidance in response to personalisation.

While customers in all groups tended to see transaction prompts as accusatory or intrusive, they would also be most likely to take notice of and act on these prompts. However, there was an increased risk with transaction prompts of unintended or undesirable responses, or behaviours. These included increased stress and worry while completing the form, particularly for less confident or experienced customers completing close to the deadline. If the amount presented was too low, there was a risk that otherwise compliant customers might act on what they see as a low-risk opportunity to underreport. Even if data is correct, customers could be encouraged to under-report if only some, but not all, sources of income are included. For example, if a figure for only one online marketplace is given when a customer sells on multiple online marketplaces.

Therefore, source prompts and those based on prior returns data provided the preferred level of personalisation and greatest potential to lead to positive customer behaviour. However, for these prompts to result in positive reported behaviours, they need to:

Come across as a genuine attempt to help: if prompts are seen as accusatory or a sign of distrust in the customer, they could lead experienced or ordinarily compliant customers to be less motivated to get their taxes right, or could cause unnecessary worry for less experienced or confident customers
Feel relevant to the individual: if prompts are not sufficiently personalised to feel tailored to the individual, they could be ignored or misunderstood as generic or not requiring action
Be informative: customers should gain from engaging with the prompt, that is, it should contain new or useful information that they are able to action

How are customer perceptions of HMRC influenced by the use of personalised digital prompts?

Customer views on the use of personalised prompts by HMRC were not straightforward. They could simultaneously see both benefits and downsides to use. Personalisation of the ITSA was felt to be consistent with the modern digital world and an indication that HMRC are keeping pace with technological developments. For customers who saw this benefit, personalised prompts provided reassurance in creating a fairer tax system and targeting those who were deliberately fraudulent. However, there were also negative implications from the use of personalised prompts:

for experienced and confident customers, personalised prompts were seen as signalling distrust in the taxpayer and introducing unnecessary and unhelpful complexity to the ITSA. These customers wanted the ITSA to remain in its current format, based on a relationship of trust between HMRC and the taxpayer to accurately declare their income
the inclusion of incorrect data posed specific risks to HMRC’s reputation. Customers explained that incorrect data would undermine trust in HMRC’s systems more generally and, if values were too high, could be seen as an attempt by HMRC to ‘over-tax’ them

Question 3

3. Introduction

Accepted Answer

3.1 Background to the research

HMRC is committed to using new technology to create a more effective tax system and help customers get their tax right first time. As part of this, HMRC is undertaking a Digital Prompts Programme which involves presenting nudges to customers as they complete their tax return, highlighting potential mistakes and directing them to relevant guidance. While prompts are regularly used by HMRC, data mastery now allows use of more personalised ‘intelligent’ prompts that draw on a range of data sources.

However, there are some associated risks involved in using personalisation, including higher administrative costs to HMRC and the risk of using incorrect data. While quantitative trials are ongoing, HMRC required a more detailed understanding of whether, and how, personalised prompts can lead to more positive behaviour than basic prompts. HMRC therefore commissioned the National Centre for Social Research (NatCen) to carry out qualitative research into customer perceptions of personalised prompts and their reported behaviours as a result of digital prompts.

3.2 Research aims

The aims of the research covered the following key themes and related research questions:

Customer understanding and capability

did customers understand the prompts?
did they understand what action was required and how to take it?
did they understand that prompts had been personalised using third party data, or data from prior returns?

Perceptions of prompts and reported implications for behaviour

what did customers perceive as the intention behind the prompts?
did they perceive them as helpful?
how did customers react to the tone, formatting, and location of prompts?
did personalised prompts lead to an increase in positive behaviours, compared to basic prompts?
were there unintended behavioural consequences of using personalised prompts, such as mis-reporting?

Perceptions of HMRC

how did the use of personalised prompts impact customers’ perceptions of the power, reach and motivations of HMRC?
how did customers feel about the use of third party data?
what were the implications if the data used was incorrect?

Although customers’ perceptions, and their reported behaviour, were originally intended to be separate themes in the research, the findings show that behaviour is closely linked to perceptions. Therefore, they are presented together in the report.

3.3 Brief overview of methods

The research involved 45 in-depth interviews with customers who had completed an online ITSA in the last 12 months without the help of a paid intermediary.

Topic guide and stimulus development

To achieve the aims of the research, the interviews used hypothetical vignettes to present customers with prompts that could be used in real time as they file their ITSA. Participants were presented with scenarios that were tailored to their own circumstances. To inform design of the vignettes and interview stimulus, a scoping workshop was held at the outset of the research to refine the research aims and HMRC’s priorities for testing. NatCen then developed, in collaboration with HMRC, interview stimuli that presented prompts with different levels of personalisation.

4 different types of prompts were developed and tested:

basic prompts with no personalisation. These presented the same generic message to all customers. For example:

HMRC routinely receives data from various sources and care is needed to ensure the property pages are completed correctly and accurately.

prior returns prompts were personalised using data held by HMRC from prior Self Assessments completed by the customer. For example:

You now rent out fewer properties than last year. If you have sold a property for more than £48,000 you may need to declare this in the Capital Gains Tax section.

source prompts were personalised using data from a third party. These prompts named a source specific to the customer but not transaction amounts. For example:

HMRC receives data on rental income and landlord properties. That data tells us about taxpayers who have rental income, both in the UK and abroad.

transaction prompts were personalised using data from a third party and included transaction amounts specific to the individual. For example:

HMRC has received data from [company] on your rented property at [address]. This data includes rental receipts of £3,000 for this period of Self Assessment.

Feedback was provided to HMRC following the first 6 interviews and final amendments made to the stimuli to ensure that it was fully meeting the objectives of the research.

Sampling and recruitment

The research used a purposive sampling approach to include key groups of interest. While the report only presents the views of those customers interviewed as part of the research, these are expected to reflect the full range and diversity of views within the population of customers who complete an ITSA online. Participants were recruited to take part in the research either through the NatCen Panel or by an external recruiter. A screener was used to determine eligibility to participate with quotas set against characteristics including gender, age, and customer type. Number of years completing an ITSA was also recorded and monitored.

The customer types chosen for the research covered the main reasons why a customer may have to complete an ITSA. These were landlords, those with unearned income (such as from a property sale), micro-businesses and gig economy workers. Table 1 presents the final achieved sample.

Table 1: Final achieved sample

Characteristic	Quota	Interviews
Gender
Male	Min 20	22
Female	Min 20	23
Age
18 - 34	Min 10	8
35 - 49	Min 10	15
50 +	Min 10	22
ITSA customer type
Landlords	Min 8	13
People with unearned income	Min 8	9
Micro-businesses	Min 8	11
Gig-economy workers	Min 8	12
Years filed an ITSA
Up to 2 years	NA	7
3 years or more	NA	38
Total	45	45

Fieldwork

Interviews were carried out in May and June 2023. Each interview lasted approximately 1 hour and was carried out using Zoom. Interviews were recorded with the participant’s consent and transcribed.

Analysis

NatCen’s framework approach was used to summarise the rich qualitative data on a cross-sectional case and theme basis. Findings under key themes related to the study aims and research questions were summarised across each interview. This enabled the researchers to look across the summarised data to identify key insights about how participants views, perceptions and behaviours varied across key characteristics.

Question 4

4. Understanding of prompts

Accepted Answer

This chapter explores customers’ understanding of personalised prompts. It begins by exploring the extent to which customers understood the basic content of prompts. It then discusses the extent to which customers understood what action was required from prompts and that prompts had been personalised.

4.1 Customer understanding of message content

Customers generally had a good understanding of the basic content of messages, although they tended to understand personalised prompts better than basic ones. This was because basic prompts were perceived as too vague or generic to be useful or meaningful. For example:

HMRC receives data from a variety of sources which is used to check the accuracy of taxpayers’ tax affairs. Please make sure all the information you provide in this return is accurate.

The use of the phrase “a variety of sources” was felt to be particularly vague. Customers described how the prompt did not provide them with any additional information they did not already know or had not already been mentioned elsewhere in the ITSA. The prompt also did not provide customers with any information that would help them or that they could act on.

While personalised prompts were generally better understood, there were features of these prompts that caused some confusion among customers, namely the use of technical language and unfamiliar terminology. Examples of such terminology included: “form SA105”, “Capital Gains Tax”, “other tax adjustments” and “used to risk-assess the tax affairs of taxpayers”. Customers were also confused by references within prompts to numbered boxes, as they recalled that the online ITSA did not number answer boxes.

If unfamiliar with the terms used, participants said they might seek further information and clarification, such as through links to further guidance provided in the prompts themselves, where applicable. Participants suggested use of technical language could result in the ITSA taking longer to complete than if more simple terminology was used or clarifications were provided within prompts.

Customers also suggested that the inclusion of some additional information within prompts could further enhance their clarity and usefulness, such as explicitly stating the Capital Gains Tax and VAT thresholds.

4.2 Customer understanding of required action

Customers generally had a good understanding of what action was required from the prompts. However, there was a distinction between customers who felt they already knew their tax obligations prior to being shown the prompts, and those who did not. For the latter group, prompts helped them understand what they were required to do.

Customers tended to be able to distinguish between prompts that indicated they may have an additional tax obligation to fulfil at a later point, such as registering for VAT, and those where they were required to take action at the section of the ITSA being presented. For more experienced and confident customers, prompts referencing additional tax obligations confirmed their awareness and understanding and served as a reminder to fulfil these obligations. Throughout the report we refer to confident and experienced customers, as the research showed that length of time completing the ITSA did not always equate to tax confidence, and vice versa.

Where participants were unsure of the tax obligations referenced in the prompts, they welcomed the inclusion of links to further guidance, as these would allow customers to understand their obligations and act on them. Customers highlighted that links to further guidance would be most helpful if they opened in a new tab, so they did not leave the ITSA form. Customers also felt it would be helpful if the links brought customers to simple and easy to understand information, for example, a calculator or short questionnaire to easily determine if an obligation applies to them.

Customers also felt that some prompts were missing necessary information which they would need to know in order to respond. For example, where sections of the ITSA form did not specify the timeframe that customers should be declaring income for, they would be unsure of the correct figures to enter when reporting turnover. In relation to the transaction prompts, participants questioned what dates the figures applied to. For example, in the prompt below, customers generally wanted to know the specific timeframe covered by the financial data.

HMRC has received data from the [bank] branch in Wexford of your account 12345678. Please ensure these account details and the taxable interest of £2,589 are both included in your return. If you omit income or make an inaccurate return, HMRC will take action to recover any additional liabilities due which may include interest and penalties. These penalties are more severe for omitted offshore income/assets.

Customers further reported challenges in understanding sections of the ITSA form itself and knowing how to respond accurately. They therefore suggested that prompts could provide further clarification on technical aspects of the form.

4.3 Customers understanding of personalisation

Customers interviewed were not always clear on the difference between basic and source prompts, and that prompts were personalised to their specific circumstances. Depending on the data source referenced, customers did not always understand the prompts were personalised and instead interpreted the messages as generic to all customers. This was particularly in relation to debit and credit card payment processing services, whereas those referring to more specific sources, such as online marketplaces or sales of second homes, were recognised as personal to the individual.

There was a good level of understanding among customer groups about how prompts related to third party data were generated and why they had been triggered. However, some customers questioned how and why HMRC had accessed their data. On the other hand, customers were less likely to spontaneously understand how prompts based on prior returns had been generated, as the data source was less clearly signposted. The below prompt highlights where the data source was less clearly signposted in prompts based on prior returns.

WARNING: Your declared turnover has been just below the VAT registration threshold for a period of time now. Please ensure you are calculating your turnover correctly before submitting this return to HMRC. A failure to do so may result in penalties. For more information on calculating your turnover, refer to the VAT registration guidelines (opens in new tab).

There was also mixed understanding among customers around answers that had been pre-selected. Without further explanation, some customers struggled to understand the concept of options being pre-selected for them by HMRC, rather than the customer selecting the option themselves. This was possibly because pre-selection within the ITSA, unlike prompts, is a new approach. If introduced, customers suggested pre-selection should be used consistently across the ITSA to avoid confusion or mistakes by customers thinking they themselves had already selected a particular option and therefore not checking it.

Question 5

5. Perceptions and implications of prompts

Accepted Answer

This chapter explores how customers felt about the different types of prompts they were presented with, and how the prompts could influence their behaviour.

5.1 Influence of personalisation on perceptions and behaviour

This section explores how the different levels of personalisation influenced how customers perceived prompts, including whether they were viewed as relevant and informative, and what customers understood the intention behind the prompts to be. How customers perceived the prompts was closely related to how they thought they would behave in response to them: whether they would ignore them or not, and if not, what exactly they felt they were most likely to do. This section therefore also explores these behavioural responses.

Influence of personalisation on perceptions of prompts

Customers perceived the prompts as useful in a variety of ways. For all customer types, prompts could be seen as a useful reminder to take extra care at certain sections of the ITSA, and as a way of guarding against accidental mistakes, especially when prompts appeared in response to customer actions. For less confident customers, and for gig economy workers who may be less certain of their exact income, prompts could be seen as an encouragement to double check their figures.

“I think it’s a very good reminder to make sure to do it properly, to declare income properly.”

(Gig economy worker, female, aged 50 to 64, 3+ years completing ITSA)

“It can be quite confusing for a self-employed or sole trader to keep all the transactions, so it would be very convenient.”

(Micro-business owner, female, aged 35 to 49, less than 2 years completing ITSA)

Despite this, across all customer groups, prompts tended to be seen as an attempt by HMRC to increase tax revenue and reduce fraud, rather than an attempt to help the customer get their taxes right. Customers’ perceptions of prompts also depended on the level of personalisation used.

Basic prompts

Basic prompts, with no personalisation, tended to be viewed as either too vague to be useful, or as an unnecessary threat. How customers perceived these prompts depended largely on how experienced or confident they were completing the ITSA form. Experienced and confident customers felt they would ignore basic prompts, viewing the message as something they already knew. These customers referenced the honesty declaration at the end of the ITSA form and felt the basic prompts did not add anything over and above this. There was frustration at the perceived duplication.

By contrast, less experienced or confident customers felt that basic prompts were a “veiled warning” or an “implied threat”. This meant basic prompts had the potential to cause worry and might cause customers to contact HMRC for clarification, by phone. Customers tended to feel that this “warning” or “threat” was unjustified, because they had not entered any information that could give HMRC reasonable cause to suspect non-compliance.

“If you put in a small number after lots of years that there had been a bigger number, I think it would be fair enough to say check it…but if she has put nothing in and there is nothing to go on, I think potentially it is a bit too harsh.”

(Gig economy worker, male, aged 18 to 34, less than 2 years completing ITSA)

Source prompts

Compared to basic prompts, source prompts were generally viewed as relevant and informative, with the potential to help customers complete their ITSA form. Customers liked that source prompts felt relevant to their personal circumstances, which made them feel more like an attempt to help than basic prompts. When source prompts were triggered in response to an error, rather than shown by default, customers felt that the prompts gently implied they may have made an error, rather than trying to catch them out. Source prompts offered additional information to help customers complete their return accurately and ensure they had considered all relevant sources of income, which was appreciated.

“Because it mentions online marketplaces and she has a business that uses online marketplaces, it feels less accusatory because it is definitely relevant.”

(Gig economy worker, male, aged 18 to 34, less than 2 years completing ITSA)

Transaction prompts

While source prompts tended to be seen as helpfully pointing out a potential error, transaction prompts tended to be seen as accusing customers of deliberate underreporting. Customers felt the inclusion of multiple personal details was potentially excessive and interpreted this as “threatening” and “scary”. There was a view that HMRC needed strong justification for using highly personalised prompts, such as only showing them where customers had repeatedly entered information incorrectly. For less experienced or confident customers, transaction prompts could cause additional worry or stress when completing the form.

Prompts using data from prior returns

Customers were generally positive about the use of data from their prior tax returns where they felt this made the ITSA quicker and easier to complete. These prompts tended to either confirm customers’ existing understanding of their obligations or prompted them to read further information. However, customers felt negatively about these prompts when used to query the accuracy of information they had entered. Again, more experienced and confident customers felt this use indicated distrust, while less experienced or confident customers suggested questioning the accuracy of information entered could cause unnecessary worries where their figures were actually correct.

“I would be very stressed. ‘What do I have to do? Do I need an accountant?’ The word ‘penalties’ seems overwhelming. Message suggests you have done something wrong before you have done something wrong.”

(Gig economy worker, female, aged 35 to 49, 3+ years completing the ITSA)

Implications of personalisation for behaviour

Broadly speaking, prompts were intended to lead to 2 main types of customer response. Firstly, for customers who were already intending to report correctly, the aim was to help them report accurately and avoid error. Secondly, for customers who were being fraudulent, the aim was to encourage them to report correctly.

Reducing error amongst customers

Most of the findings from this research in relation to customer behaviour relate to the first aim of the prompts: to help customers report accurately and avoid error. It is important to be clear about the limits of the research method used. Customers were shown scenarios and asked how they thought they would respond, which can differ from how customers actually respond when faced with a real situation. A range of interviewing techniques were used to encourage participants to fully reflect on whether they would really change their behaviour in response to prompts. Nonetheless, this section should be read as a summary of the likely or potential behavioural impacts of prompts, rather than as conclusive evidence that certain behaviours would be exhibited.

In line with customer perceptions of prompts, there were differences between more and less experienced or confident customers. More experienced and confident customers were generally unlikely to take any action in response to basic or source prompts, and typically assumed that they already understood their obligations correctly. Less experienced and confident customers were more likely to feel they would act in response to these prompts, typically in one or more of the following ways: following links to, and reading, additional guidance; checking their figures were correct; or contacting HMRC to clarify their understanding of their obligations. Unsurprisingly, customers were more likely to read additional guidance when there were direct links to this guidance in the prompts.

For transaction prompts, while customers tended to see these as accusatory and threatening, they also felt they would be more likely to take notice of and act on them compared to basic and source prompts. Customers felt this was because transaction prompts were significantly different from what they were used to seeing in the ITSA, and they were much more likely to have an emotional response to this level of personalisation. However, it is important to note that although transaction prompts may be more likely to lead to positive behaviour, such as checking figures or reading guidance, there were also a range of unintended or negative behavioural responses mentioned by customers.

Discouraging fraudulent reporting

It is important to note that none of the customers interviewed for this research stated that they had been fraudulent when completing their ITSA forms in the past. As such, it is not possible to comment with confidence on whether the prompts would be likely to prevent fraudulent reporting. Customers believed that transaction prompts were the most likely to cause fraudulent customers to rethink their actions, given both the tone and formatting of the prompts, and what they revealed about HMRC access to data. However, customers generally felt that even these prompts would not be sufficient to dissuade deliberately fraudulent customers from underreporting.

“Someone who sets out to purposely defraud the tax office is not going to be scared off by a bit of strong language and some red writing.”

(Gig economy worker, male, aged 50 to 64, 3+ years completing the ITSA)

5.2 Customer response to pre-selected options

In general, customers responded positively to pre-selection when it was seen as making the ITSA easier to complete. Pre-selection was felt by some to reduce customer decision making and generally lower the cognitive load.

“They are starting where [the customer] finished last year so this is helpful”

(Landlord, male, aged 65+, 3+ years completing ITSA)

“It is showing that they [HMRC] know this information and it is guiding them to make the right choices”

(Landlord, female, aged 35 to 49, 3+ years completing ITSA)

However, different customer groups reacted differently to pre-selection. Landlords, and those with unearned income, felt that pre-selection was useful in guiding them through the ITSA, but felt that it should be applied consistently across the form

“To be helpful it has to be used throughout the form, otherwise I might think I have already answered”

(Landlord, male, aged 65+, 3+ years completing ITSA)

In contrast, micro-business owners and those working in the gig economy did not feel as positive about pre-selection and preferred to have control over the form themselves. Micro-business owners, for example, expressed concerns about previous tax years being used to inform pre-selection, because their circumstances could change significantly from year to year:

“It would be very helpful, but then when the business changes or they diversify… you may miss out on some things if you were changing the business.”

(Gig economy worker, female, aged 50 to 64, 3+ years completing ITSA)

5.3 Customer response to tone, formatting, and presentation

Across all prompt types used in the research, the tone, formatting, and presentation of prompts influenced how customers perceived them.

Tone of prompts

Customers responded negatively where they felt the tone of prompts indicated wrongdoing, even when customers had entered information correctly. Customers expressed they would feel increasingly stressed with the process of filing an ITSA. The following prompt, for example, was felt to have an accusatory tone:

WARNING: Your declared turnover has been just below the VAT registration threshold for a period of time now. Please ensure you are calculating your turnover correctly before submitting this return to HMRC. A failure to do so may result in penalties. For more information on calculating your turnover, refer to the VAT registration guidelines (opens in new tab).

“I would be very stressed, what do I have to do? Do I need an accountant? Word penalties seems overwhelming. Message suggests you have done something wrong before you have done something wrong”.

(Gig-economy worker, female, aged 35 to 49, 3+ years completing the ITSA)

Customers instead preferred prompts which were framed as information to help them complete their ITSA correctly and meet their tax obligations. The use of help links and guidance within the prompts was also welcomed. The following prompt is a good example of this:

IMPORTANT: Your declared turnover is approaching the VAT registration threshold. We want to help you understand your VAT obligations and help you get it right. For more information on when/ if you need to register for VAT, refer to the VAT registration guidelines (opens in new tab).

“There was guidance and help and it felt like this prompt was guiding [[the taxpayer]] to complete the form correctly”.

(Landlord, female, aged 35 to 49)

Prompts were perceived as informative or helpful when they used second person narrative. Customers perceived the use of ‘you’ rather than ‘taxpayers’ within the prompts as directly addressing them, which made the content and the action required in the prompts feel more relevant:

“Use of direct language is important…we have info about you… the message would not feel generic or vague.”

(Landlord, female, aged 35 to 49, 3+ years completing the ITSA)

Formatting of prompts

The visual presentation of prompts influenced how customers reacted to the prompts. Customers were shown prompts in black, red and blue text. Participants perceived the black text as being neutral and standard for official forms, and easy to read. In comparison, red prompts were immediately interpreted as a warning and an indication of wrongdoing: the red text “felt accusatory and made me panic”. Customers perceived the blue text as being calming, informative and “less ominous”.

Presentation of prompts

The location of the prompts on the screen influenced how customers reacted to them. When prompts appeared above the form, they were perceived as providing generic information:

“If it’s at the top it’s bolder and more of a reminder that we are in the process of checking up on lots of people, so more of a warning to lots of people.”

(Micro-business owner, female, aged 40 to 64, 3+ years completing the ITSA)

When prompts appeared below the form, they were perceived as being a specific warning or appearing as an ‘error’ message to alert them that something they had entered was incorrect. If the purpose of the prompts was to prevent errors, some customers felt presenting the prompt below the form might be more effective. Ultimately, customers felt the placement of prompts on the ITSA form should be consistent throughout, so customers knew where to look for information. If the message was to move, customers were concerned they may inadvertently miss the message.

5.4 Customer response to incorrect data

Customer reactions to incorrect data within the prompts depended on their experience and level of confidence in completing the ITSA.

More experienced or confident customers

When more experienced or confident customers were presented with incorrect data, they expressed their frustration at the added complexity of having to deal with the error. These customers felt that they would be responsible for contacting either HMRC or the source of the incorrect data (for example, a third party) to ensure it was corrected:

“It would be a bit of a pain in the neck…Ideally you would want to contact HMRC to say, “the figures are wrong you need to get in contact with [the third party platform]”. The onus should not be on the individual as self-employed people are busy.”

(Micro-business owner, female, aged 40 to 64, 3+ years completing the ITSA)

Instead, these customers wanted to be able to easily override or report incorrect data presented in the ITSA. Customers wanted to be able to provide evidence for the changes within the form itself by submitting proof or information from their own records in the form of receipts and invoices. Customers felt this would ensure they were not contacted by HMRC after filing the ITSA:

“If info was incorrect, I would like some kind of button that could be clicked… where there’s a free-form box, and I can upload evidence to support the dispute. It needs to be something like that because God save us from having to call HMRC. This is an online form so the ability to dispute it should be available on the form.”

(Gig economy worker, male, aged 65+, 3+ years completing the ITSA)

Less experienced or confident customers

When less experienced or confident customers were presented with incorrect data, they felt it would lead to increased stress and worry when completing the ITSA. For these customers, a message telling them the information they had entered was wrong would leave them feeling “accused or chased up when they haven’t done anything intentionally wrong” (Gig economy worker, female, aged 18 to 34, 3+ years completing the ITSA).

For these customers, after being presented with incorrect data, they would initially re-check their own records and then would either contact the relevant third party company or HMRC directly for reassurance. If the incorrect data was provided by online marketplaces or letting platforms, these customers would be more likely to contact these third parties directly than HMRC. However, if the incorrect data presented was received from card payment services (such as card transactions), these customers would be more likely to contact HMRC in the first instance. Customers expressed that if they had multiple accounts it would be easier to contact just HMRC than multiple card payment services.

Customers felt that undertaking these additional checks would likely lengthen the time needed to complete the ITSA. For customers who completed closer to the deadline, the risks included late filing and the associated fees:

“If information was wrong, this would prevent them from submitting the form. You have to sign something to say the information in this form is accurate. If this is wrong you cannot sign it and you cannot submit.”

(Landlord, male, aged 65+, 3+ years completing the ITSA)

5.5 Potential for prompts to lead to underreporting

The research suggests that transaction prompts could have negative behavioural consequences, including underreporting. Customers identified 3 possible ways in which incorrect reporting could arise as a result of personalisation:

if HMRC presents incorrect amounts, which are too low, customers could be more likely to use these amounts
customers may infer that the data source in the prompt is the only data HMRC holds and underreport elsewhere on the ITSA
customers may be confused about the relevant income to report and make mistakes

Underreporting as a result of incorrect data

Customers felt that when presented with incorrect data, those who would otherwise be compliant may act on what they see as a low-risk opportunity to underreport. Customers may feel that as the information HMRC holds on them is incorrect, they are unlikely to be caught and if they were to be caught, they could then use HMRC’s mistake as an excuse if accused of wrongdoing later. This view was more prominent for those customers who felt that HMRC were using data to “try and catch them out”.

Potential for misreporting in response to data sources

Where data presented in the prompts was correct, customers felt taxpayers may be encouraged to underreport if they thought the data source referred to in the prompt was the only data that HMRC had access to. This particularly applied to customers who had multiple sources of income or who received payment through more than one platform, for example, card and cash payments. As such, presenting data on card transactions appeared to confirm customers’ beliefs that HMRC could not verify or check cash payments, which may in turn lead to mis-reporting or under-reporting as customers believed “there would be no record of cash payments”:

“I think there is a problem that for instance the government did not highlight a certain site which I also participate in, then I would think I could use that as an excuse. If the government brings it up later on, I could use that as an excuse to say that you didn’t tell me that and I forgot about it because it was not highlighted.”

(Gig economy worker, female, aged 50 to 64, 3+ years completing the ITSA)

Potential for misreporting in response to relevant sources of income

Where customers had multiple sources of income the inclusion of some data, but not all, could cause confusion about the relevant income to report. For landlords, there was debate about whether they should declare any income they received from informally renting to family and friends within the ITSA. For other customers who had multiple sources of income, for example income from an online and a physical shop, there was uncertainty about whether a separate ITSA should be completed for each source or if the figures should be combined:

“There is no way you can get this information if family and friends are going there and paying cash, there is no way HMRC would know about this”

(Landlord, male, aged 65+, 3+ years completing the ITSA)

5.6 Features of prompts perceived to be associated with positive behaviour

Given all of the above considerations, customers generally felt that source prompts were more likely than other prompts to lead to positive behaviour. However, customers felt that in order to create these positive behaviours, these prompts should be:

presented as a genuine attempt to help customers. If prompts were seen as accusatory and indicative of a lack of trust, this could lead experienced or ordinarily compliant customers to be less motivated to get their taxes right than they otherwise would have been, and could lead to less experienced or confident customers feeling significant worry and stress
relevant. If prompts did not have enough personalisation to feel relevant to the customer, they were more likely to be ignored or interpreted as a generic message that did not require any action
informative. Customers should stand to gain by engaging with a prompt: if a prompt did not contain information that customers were able to action, it was not seen as useful

In order to be perceived as a genuine attempt to help customers, relevant, and informative, there were several features of prompts that were important. Prompts should:

include sufficient personalisation to feel tailored to individual circumstances
include links to further guidance
use data from prior returns rather than third parties, where possible
be triggered in response to customer actions
provide reassurance to customers where they are completing the ITSA accurately
address the customer directly through the use of second-person narrative, for example, “Records show that you may…”
use plain English
use a helpful and support tone, rather than indicating wrongdoing
use friendly formatting, for example, avoiding red text

Question 6

6. Perceptions of HMRC and use of personalised prompts

Accepted Answer

This chapter explores the impact that personalised prompts had on customer perceptions of HMRC.

6.1 Customer attitudes towards HMRC’s access to and use of third party data

Customer views on HMRC’s access to and use of third party data were not straightforward. Generally, they supported the use of data to help customers complete an ITSA accurately, but still held concerns about data sharing, particularly by private sector and overseas companies.

Customers supported the use of data in personalised prompts for a range of reasons. These reasons were as follows:

micro-business owners viewed data sharing with online marketplaces as no different to an employer sharing financial information with HMRC through PAYE
older and more experienced customers expected, or were already aware, that HMRC had access to a large amount of information about customers. This was particularly the case with data that was already held by government (such as from previous tax returns and the land registry). There was a view that customers should assume that HMRC already have this information about them
there was a view that personalised data could help customers complete their returns accurately and ensure that customers are paying the right amount of tax

“It could result in a fairer tax system if those who were deliberately dishonest were targeted.”

(Landlord, female, aged 35 to 49, 3+ years completing the ITSA)

Micro-businesses and gig-economy customers viewed data sharing as keeping pace with modern technology and using technology to aid in the process of completing an ITSA. It was also viewed as a step towards Making Tax Digital, which was welcomed by experienced customers.

Although generally customers supported the use of data in personalised prompts, some were strongly opposed to HMRC’s access to and use of data (particularly third party data) and others, though supportive overall, still expressed concerns. Concerns related to data security, informed consent, privacy, data accuracy, and varied by the type of data used in prompts.

Data security

Customers were concerned over the security of their data and how protected it was against data leaks, hacking and being shared with other organisations. Customers suggested that they may be supportive of the use of third party data as long as all data protection rules and legislation are followed and data protection systems in place function properly.

Customers expressed concern over third party data being shared without the customer’s knowledge or informed consent. They explained how including information on this data sharing agreement in the third party’s terms and conditions was not sufficient, given that most people do not read these thoroughly. Instead, customers suggested that HMRC should provide them with further information around how the data sharing complies with legislation and that customers should be told more explicitly by third party platforms that their data would be shared and used in this manner. For example, it was suggested that a note could appear next to information on how much a customer has earned on the platform:

“Maybe where it says how much you’ve earnt or your income, a little note there saying this information does go to HMRC.”

(Gig economy worker, female, aged 18 to 34, 3+ years completing the ITSA)

Privacy

HMRC’s ability to access data from third parties raised concerns among customers around privacy. Customers expressed concern over how much the government knows about the public and emphasised how they believed in the right to privacy. HMRC accessing and using data from third parties was viewed as intrusive, with customers likening HMRC to “Big Brother”.

Data accuracy

Customers were concerned about the accuracy of the data that third parties would share with HMRC and what the consequences would be for the customer should the data be inaccurate. It was suggested that should third party data be accessed and used to personalise prompts, third parties would need to take ownership to ensure the data being shared was accurate. The ability to check the data easily, should any discrepancies arise, was further highlighted. Customers explained how HMRC would need to be open to the idea that data supplied by third parties may be inaccurate, and the first response to any discrepancies should not be accusing customers of wrongdoing.

Type of data

Attitudes towards HMRC’s access to and use of data also varied by the type of data used to personalise prompts. Customers were more supportive of data that was already held by government, such as data from previous tax returns and the Land Registry. As HMRC already have access to this data, customers viewed its use as justifiable and less intrusive. In comparison, customers were more concerned about the use of data from private companies and overseas sources, such as banks or letting agents. Customers felt that HMRC should not have access to data from overseas for a variety of reasons, including:

the perception that HMRC do not have jurisdiction to access information from outside the United Kingdom
the perception that such data should be used instead by the government in the country where the income was generated to personalise tax paid in that country
the perception that the use of overseas data could lead to greater inaccuracies. Customers questioned how HMRC would know for certain that money generated from overseas was income that they needed to pay tax on. Customers appeared to believe that HMRC would not have access to detailed and specific information from overseas and questioned how HMRC would know that any income from overseas was not a gift, for example

The concerns that customers had over HMRC’s access to, and use of, data tended not to change their perceptions of HMRC. Instead, they made customers feel more negatively about, and question, the private companies that data was sourced from. It was suggested these concerns could lead other customers to avoid using certain sites, like online marketplaces.

6.2 The impact of personalised prompts on customer views of HMRC’s reach and power

The use of personalised prompts tended to have no effect, or a positive effect, on customers’ perception of HMRC’s reach and power. Negative views about data sharing did not tend to influence perceptions of HMRC, because customers tended to feel negatively towards third party platforms instead.

Positive or limited impact on perceptions

Customers reported that personalised prompts either had no effect or a positive effect on their overall view of HMRC. Reasons given for this were that:

prompts were simple reminders to not make false claims
these types of messages from HMRC were expected
customers already viewed HMRC negatively (meaning the introduction of prompts confirmed rather than changed their existing perception)
prompts were appropriate for HMRC to use. This was because the public were required to pay tax and it was HMRC’s role to ensure the correct amount of tax was collected and to reduce fraudulent reporting
prompts could make the process of completing an ITSA simpler for customers and help them complete it more accurately and quickly
prompts could improve customer views on HMRC competency, by increasing confidence that HMRC are collecting tax owed

“[The prompt] gives you the idea that HMRC are monitoring and keeping up to date with stuff, they are competent at collecting tax, puts more confidence in what they are able to do”

(Landlord, male, aged 18 to 34, less than 2 years completing the ITSA)

Negative impact on perceptions

However, other customers reported that the use of personalised prompts would lead them to view HMRC more negatively. For example, customers noted that they would perceive HMRC as invasive, untrusting of taxpayers and “money-pinching” (although customers did acknowledge that it was HMRC’s role to collect all tax owed).

Different types of prompts were also found to elicit different perceptions of HMRC. For example, basic prompts led to customers viewing HMRC as threatening, rather than supportive, while transaction prompts made HMRC seem aggressive, accusatory, and untrusting of customers. Similarly, where data from prior returns was used to query accuracy, this was felt to indicate distrust of the taxpayer.

Inaccurate third party data

Customers also discussed how their perceptions of HMRC would be negatively impacted if third party data used in personalised prompts was inaccurate. They described how it could make customers lose confidence in HMRC’s competency and lead customers to question the quality and accuracy of wider HMRC records.

Cookies on GOV.UK