Coronavirus and the experiences of UK businesses, textual analysis: March 2020 to July 2020

This article presents qualitative responses for the first time from the Business Impact of Coronavirus (COVID-19) Survey (BICS) for a select number of questions from Waves 2 to 9. We focus on those covering turnover, international trade, business experience of government schemes and business working arrangements. These qualitative responses vary across sector, business size and over time and help provide some further anecdotal evidence on how these businesses have been impacted by the shocks to the economy. This analysis summarises the most commonly cited words that businesses responded with, in the context of popular phrases, and establishes how these have changed in frequency over time to gain insight into qualitative messages behind the BICS responses.

Table 2 in Section 9 shows the relevant questions and how, if at all, they have changed across Waves 2 to 9.

• Business comments describing what had specifically affected turnover in relation to the top word "closed" generally related to "downstream" business closures, where businesses had been affected by either their own business, or businesses they supply, closing.

• Of businesses responding to international trading questions, in general the prevalence of comments relating to the impact of supply related logistical challenges and customer related issues impacting trade have slightly declined from Wave 2 to Wave 9.

• From Wave 6, business responses from those who have staff under "other" working arrangements have generally been dominated by comments relating to authorised parental leave.

• Of businesses who had applied for at least one government scheme from Wave 3 onwards, comments indicated that businesses consistently found the application process at least relatively easy and these comments were extremely more prevalent across waves than those comments indicating that their experience was difficult.

• Business comments relating to descriptions of other government support that could help their business were consistently dominated by businesses commenting on new and industry specific measures of support.

For the purpose of this analysis, common phrases have been identified by analysing the most frequently written single words in response to each question. From these single words, a common phrase has been derived based on variations of instances in which the top word appeared. Table 1 in Section 7 provides a detailed outline of the common phrases that we have identified for each question we have analysed along with further information on the methodology.

The following sections present the most common phrases in the form of a time series of percentages of each common phrase, relative to the total number of comments provided to these questions in each wave. A detailed breakdown by industry of the total number of comment counts, total counts of common phrases, and percentages over time, can be found in the detailed dataset in Section 6.

Turnover

Across all waves, businesses have been able to provide additional information on how their turnover has been affected.

By sifting all qualitative responses that included the top word, "closed", in Wave 2, the two main common phrases that were established from numerous variations were:

• "closed upstream"
• "closed downstream"

The former reflects all instances in which businesses commented that the businesses that they received supplies from had experienced some form of closure, while the latter reflects all instances where businesses indicated in their responses that their own business, or businesses they supplied, had closed. These include industry specific closures, partial closures and cancellations through the supply chain.

In Wave 2, the majority of businesses who included the word "closed" in comments relating to turnover, indicated that either their own, or supplying businesses, had experienced some form of closure that affected their business turnover. The common phrase "closed-downstream" encompassed most instances in which the word "closed" appeared in comments. This emphasises the impact on responding businesses operational capacity at an individual business level, and the reliance on consumer demand on turnover.

By Wave 9, the frequency in which the word "closed" appeared was significantly lower than Wave 2, and from Wave 5 onwards, steadily began decreasing in business comments relating to "closed-downstream" circumstances. As the comment "closed" has reduced over time this is also shown in the Coronavirus and the economic impacts on the UK bulletins where there has been an increase in the number of businesses re-opening, and the change in percentage of businesses temporarily paused trading has fallen from 24% in Wave 2 to 7% in Wave 9.

At an industry level, the accommodation and food service activities sector reported the highest percentage of the word "closed" in Wave 2 in respect to all those that answered the question in this industry, at 68%, compared with 23% in Wave 9. In Wave 9, the sector with the highest frequency of comments including the word "closed" in some way was the arts, entertainment and recreation sector, at 38%. In comparison, the professional, scientific and technical activities sector reported the lowest percentage of comments including the word "closed" in Wave 2, at 25%, decreasing to 11% in Wave 9.

The percentage of businesses reporting comments on closures in earlier waves will reflect the impact of the UK lockdown imposed on the 23 March 2020, and the introduction of easing of lockdown restrictions from 23 June 2020 will likely explain the reduction in frequency of the word "closed" in later waves. As businesses began re-opening from June, and trading re-commenced, the way in which closures were affecting turnover will have reduced.

International trade

Business Impact of Coronavirus (COVID-19) Survey (BICS) has continuously asked businesses in detail about how their exporting and importing has been affected since the start of the coronavirus (COVID-19) pandemic.

For the export specific question, sifting all qualitative responses that included the top word "export" and aggregating the numerous variations in which it occurred, the most common phrase that was tracked was "unable to export". This phrase encompasses a variety of responses that businesses indicated in some way that exporting had been stopped or that the business was unable to export.

The second top word that appeared in businesses responses to the export question was "customer"and have been captured under the following two phrases:

• "customer closed"
• "customer country lockdown"

The former reflects instances where businesses have indicated that customers have closed or where they have delayed or stopped, while the latter refers to all comments that included the word customer and described that they had been impacted by government and/or global lockdown restrictions.

The slight decline in the percentages of comments from businesses reporting that they were unable to export from Wave 5 onwards may reflect the re-opening of export destination countries previously in lockdown. In general, the percentage of comments indicating businesses were unable to export has remained fairly stable across waves.

Figure 3 also shows that comments from businesses that included customer related problems in relation to exporting, were higher in the initial waves than in latter waves of BICS. From Wave 4 onwards, customer related issues in exporting have flattened off and have remained quite consistent up to Wave 9.

Customer closures and global lockdown restrictions have dominated the majority of "customer" related comments and signified comments emphasising the impact of global lockdown restrictions at the peak of the coronavirus (COVID-19) outbreak - which have since been eased gradually from Wave 3.

For the import specific question, the common phrase was "goods logistics restrictions". This encompasses comments in which businesses indicated the importing of "goods" had been delayed, restricted or stopped in some way.

The second top word that appeared in businesses response to importing was "supply / suppliers"and has been captured under the phrase "supply chain challenges". This encompasses all instances where businesses have indicated that suppliers have closed or that importing had been affected based on supply chain prices.

From earlier waves of BICS, there has been a gradual decline in the frequency of business comments relating to "goods" logistics restrictions and "goods" not being required as having affected imports.

In relation to supply related comments, the percentage in which the word "supply" has been included in comments relating to imports has declined from earlier waves of BICS. From Wave 4 onwards, there has been no major change in the frequency in which the two common phrases occurred across comments. The initial strength of comments attributing to supply issues affecting importing capability may reflect the immediate reaction to COVID-19 and hence the business adaptation from this Wave onwards. 

The routing of the questionnaire in relation to international trading questions has changed over iterations over BICS, and between Wave 3 and 4 the population that answered this question changed along with the wording of the question. This, along with sample variation based on the voluntary nature of the survey, may partially explain variation in the percentage of comments including the common phrases.

A detailed breakdown, split by industry, wave and common phrase can be found in Section 6. Caution should be taken for those industries with low response counts and results may not reflect all exporting and importing business experiences.

Government Schemes

From Wave 3, following the announcement of the government's Coronavirus Job Retention Scheme (CJRS) and other additional business support schemes, a free text question was added to provide insight into businesses experience of applying to these schemes.

Subsequently, two main common phrases were derived from the top word, "easy", to distinguish between businesses who found the application process relatively or extremely easy. For the purpose of this analysis, responses have not been filtered by government scheme and so should be treated with caution as not all businesses would have applied to all government schemes on offer.

We have also looked at the percentage of businesses commenting variations of the word "difficult", to polarise the responses and establish if any industry in particular was reporting a higher frequency of businesses finding applications difficult.

Figure 5 shows the percentage of comments in which businesses stated that the application process to government schemes was relatively easy, if not extremely easy, has remained steady from Wave 3. This emphasises that the experience of applying to these schemes have been straightforward for businesses across waves. Compared with the number of comments that indicated businesses had difficulty in applying for the government schemes, across all waves, the percentages of comments that alluded to this was markedly lower than those indicating it was easy.

As new business support schemes have been introduced since the initial CJRS, comments may be scheme specific and further insight into individual scheme experiences stated explicitly in comments may provide the basis for further textual analysis. This may also provide some explanation into the change in frequency of comments referring to schemes being "very easy" from Wave 7.

Across all industries, the accommodation and food service activities sector reported the highest frequency of comments relating to an "easy experience" in Wave 3, at 22%, which remained relatively stable across all waves (21% in Wave 9). At the same time, the accommodation and food service activities sector reported the highest frequency of comments relating to a "difficult" experience in Wave 3, at 20%, falling to 17% in Wave 9. The arts, entertainment and recreation sector had the largest percentage of comments in Wave 9 indicating a "difficult" experience in applying for the government initiatives, at 21% which compared with 18% in Wave 2.

In contrast, the information and communication sector, and the professional, scientific and technical services sector were consistently the sectors across all waves with the lowest percentage of comments indicating that they had a "difficult" experience applying for government schemes. Users should look to the percentages published in BICS across each wave of the percentage of businesses applying to government schemes to put these percentages into context. 

Figure 6 shows the percentage of comments across the four main phrases derived from the top word "support". These were in answer to the question asking businesses to state other support options they could benefit from, and the percentages have generally remained stable across all waves.

Across all comments including the word "support", the most common responses that included the word "support" were those related to "additional support" suggestions. This remained consistent across all waves. These comments included responses that explicitly stated increasing support options and/or industry specific support options. The second most frequent phrase, of which remained consistent across all waves included comments that indicated that they had enough support from the government.

The accommodation and food service activities sector consistently had the highest percentage of business comments that included responses on "additional support", in Wave 3, at 31%, compared with 23% in Wave 9. The sector with the lowest percentage of comments relating to "additional support" in Wave 3 was professional, scientific and technical activities sector (12%) - this changed to the manufacturing sector in Wave 9 at 18%.

Workforce

From Wave 6 onwards BICS collected qualitative responses from businesses who specified they had a proportion of their workforce classified as being under "other" working arrangements. Comments excluded all options that were available for businesses to select in the previous multiple-choice question, which included staff: working at normal place of work; working remotely; off sick with COVID-19; on furlough leave; or made redundant.

The top word which businesses included when reporting on the "other" working arrangements was the word "leave" and have been operationalised into four main categories of "leave" options. For the purpose of this analysis, we have also included here comments where businesses have just written the accompanying leave category in their comment without the word "leave" to capture all instances of "other" working arrangements. For example, if a business has just put "maternity" in the comments box, this would be included in the "parental leave" phrase even though the word leave does not appear in the comment.

Percentages have been calculated by dividing the number of times that the common phrase appeared in each wave by the number of total responses to the question in each respective wave.

From Wave 6 onwards, the most frequent comment in which businesses provided additional information on the working arrangements of “other” staff related to those on "parental leave". This remained consistent across all relevant waves and was much more frequent than the other three leave phrases.

Of those comments including the alternative categories of leave identified, the frequency of these comments has remained stable across waves and reflects the continued long-term nature of these response options. Reference should be made to the proportions of the listed working arrangements that businesses could respond to each Wave to place these working arrangements into context of the entire workforce which are published in the Business Impact of COVID-19 Survey (BICS) results.

Figure 8 presents the most commonly cited words that appear together in a pair from Waves 2, 6 and 9 in response to the question that asks businesses to explain why their turnover had been affected.

We have focused the two-word comment analysis on responses just to the turnover question, but further analysis in this article extends to other qualitative questions in relation to the most common single words. For the two-word comments in which context is required to infer more robust meaning, for example “last year” and “many customers”, further analysis is required to establish the sentiment behind these comments and will provide the basis for future investigation.

Figure 8: The COVID-19 outbreak, the subsequent UK lockdown restrictions and business closures have consistently been stated most frequently in two-word textual comments explaining turnover effects

Most common two-word comments from businesses responding to the question “Please explain in more detail how the coronavirus (COVID-19) pandemic affected your businesses’ turnover in the last two weeks”, Waves 2, 6 and 9

Notes:

1. "Two-word" common phrases have been ranked in order of the frequency of their appearance in comments in each wave with no skips between the words.
2. Comments where businesses have used numbers to support their comments have been removed so results just look at textual responses.
3. From Wave 6 the wording of the question changed to; "Please explain in more detail how the coronavirus (COVID-19) pandemic affected your businesses' turnover in the last two weeks".
4. Two-word phrases with a '-' in waves reflects when the two-word phrase was not written anywhere across all business responses to the question.
5. Wave 2 was 23 March to 5 April; Wave 6 was 18 May to 31 May; and Wave 9 was 29 June to 12 July 2020.

Download the data

Results demonstrate how business responses, in relation to the comments used to describe the reasons for their turnover being affected, have changed since the peak of the Coronavirus (COVID-19) outbreak.

Phrases that do not occur in earlier waves but increase in the ranking order, for example, “still-closed” and “re-opened”, reflect how the nature of businesses responses changed to the question and what factors were dominating reasons for turnover being affected in later waves based on the changing circumstances.

While the above figure does not represent the proportion of comments responding to each top two-word phrase, and response rates to BICS have varied over time, the sentiment of comments has moved away from explicitly attributing customer closures (ranked 3 in Wave 2 compared to a rank of 33 in Wave 9) as being responsible for the effect on their turnover.

Through looking at businesses' individual comments to qualitative, free text response options, we have been able to gather further insight into the direct experiences of businesses, to an industry level, on the impact of COVID-19. These qualitative textual response options allow businesses to explain in depth how, what and who is affecting their business and comments can be used to provide further indications of business response plans.

The continued reference to supply and customer demand across questions, and waves, emphasises the impact that disruption to procurement supply chains has on consumption and demand and reiterates the impact that closure of businesses, reduced supply and businesses cashflow has had as a consequence of COVID-19.

Where businesses provide meaningful comments to these questions, further development of data sources on the economic impact of COVID-19 can be collected and hence provide users with a more comprehensive summary on the business impact.

For further information, Business Impact of COVID-19 (BICS) survey results datasets and wave questions are available.

The sampling frame has been reviewed and refreshed across the Business Impact of Coronavirus (COVID-19) Survey (BICS) waves and since Wave 7 the sample has an increased coverage of smaller-sized businesses. For the purposes of this article, all responding businesses in Waves 2 to 9 have been taken into consideration and the article does not track the same panel of businesses over time.

By removing plain English words (such as "the" and "we") from businesses responses, we have been able to create a frequency count of the top words with meaningful sentiment from the first wave that the question appeared and subsequently create a "phrase bank" of common phrases that have been tracked across the waves. For example, the top word for the exporting question was "export", however without context there is little insight into business responses to this question, so comments are grouped under a common phrase that include the word "export" to provide meaningful sentiment.

We have tracked these phrases up to Wave 9 to establish if the prevalence of each phrase has changed over time and examine if the most common phrases have remained as prevalent in forthcoming waves.

Table 1 shows the top single words for each textual analysis question and the derived common phrase(s) for each question.

Notes:

1. An additional qualitative question was asked up until Wave 6 relating to businesses turnover – we have chosen to exclude analysis of responses to this question based on the low response rate and the fact that it was deleted from Wave 6 onwards.
2. The phrase bank phrases have been derived based on textual coding of instances in which the top word appeared and aims to capture variations of the most common business responses including the top word.
3. Each common phrase has been derived from businesses response that include the top word at least somewhere in their response.
4. The top words for turnover, international trade and prices questions have been taken from business responses from Wave 2, top words for government schemes questions have been taken from Wave 3 and top words for the workforce question have been taken from Wave 6.
5. Plurals, synonyms and alternative versions of the top words have been incorporated when tracking the top word over time to ensure all response are covered; for example, Close, Closed and Closure.

The occurrence of a word or phrase in relation to the percentage that they occur over a given time period is presented across all the waves the question was used. From this, we can observe the strength in which the phrase is prevalent in each wave to examine if the prevalence of the comment has changed over time.

The analysis in this article focuses on the percentage of the identified common phrases for each question from Wave 2 up to Wave 9.

It is important to note that every relevant comment can be counted in multiple common phrases derived for the same question. For example, businesses may have provided detailed responses that fall under two different common phrases in the same comment and so the comment is counted twice, for each phrase.

Further breakdowns on the frequency of these common phrases, split by industry, can be found in the detailed dataset in Section 6.

It should be noted that we have only derived the tracked common phrase(s) in a phrase bank from the first wave that the question appeared and so analysis does not look at "new" phrases that may have appeared in later waves. Analysis is also not representative of the same individual businesses responses over time, and instead focuses on phrase frequency relative to the total response rate for each question.

Each common phrase has been derived from businesses responses that include the top word at least somewhere in their response. Plurals, synonyms and alternative versions of the top words have been incorporated when tracking the top word over time to ensure all response are covered; for example, "Close", "Closed" and "Closure" are treated as the same word for this analysis.

Caution must also be taken when interpreting these results as a small number of responses where businesses have provided no comments at all, no meaningful comments or have responded stating reference to previous comments have not been removed. Therefore, reference should always be made to the total response for each question when interpreting these results. This is available in the detailed dataset in section 6.

Notes:

1. N/A represents where the question did not exist in the Wave
2. It should be noted that other qualitative questions exist across waves, but these have not been included here for presentational purposes.
3. Caution should be taken when interpreting results based on the changes in routing of the questionnaire and the total number of businesses that could answer the free text response questions. Further information can be found in the dataset in Section 6.