Integration testing is common across software teams, but the numbers vary with the population surveyed and the definition used. Recent and historical surveys place reported usage between 37% and 82%, while individual studies report adoption above 60% in several developer populations. Automation and ownership show similar variation.
Table of contents
- How common integration testing is
- Automation and test portfolios
- Integration testing in delivery pipelines
- Who owns integration testing
- Why survey results vary
- What the figures show
How common integration testing is
Several surveys indicate that integration testing is a normal part of modern software work. In the State of Frontend 2022, 61.5% of respondents said they had written integration tests. Solita’s developer community survey reported a similar result: 63% said their projects included automated integration tests during the two years covered by the 2022 survey. These figures describe different populations, so they should be read as parallel measurements rather than a single industry-wide rate. (The State of Frontend 2022; Solita Developer Survey 2022 - summary of results)
The Sauce Labs and DZone survey also measured integration testing among people who used automated testing. In that population, 77% reported automated integration tests in 2018, rising slightly to 78% in 2019. Because the denominator was automated-testing users, these results are not directly comparable with a survey of all developers. (DZone Guide to Automated Testing: Sauce Labs Spotlight)
Recorded test inventories show that adoption does not mean integration tests dominate every suite. In GitLab’s May 1, 2019 inventory, integration tests represented 18.2% of Community Edition’s recorded tests, 17.2% of Enterprise Edition’s tests, and 17.9% of the combined inventory. These are proportions of a specific product’s recorded tests, not proportions of teams using integration testing. (Testing levels)
| Measure | Reported result | Population or scope | Period |
|---|---|---|---|
| Integration tests written | 61.5% | State of Frontend respondents | 2022 |
| Projects with automated integration tests | 63% | Solita developer community | 2022, covering the past two years |
| Automated integration tests | 77% | Automated-testing users in Sauce Labs/DZone | 2018 |
| Automated integration tests | 78% | Automated-testing users in Sauce Labs/DZone | 2019 |
| Integration share of test inventory | 17.9% | Combined GitLab Community and Enterprise inventory | May 1, 2019 |
Automation and test portfolios
The 2025 IT Community survey places integration tests alongside acceptance tests as the most frequently selected automated-test categories. Among respondents who write automated tests, 66.3% selected acceptance tests and 66.3% selected integration tests. The same survey found 48.0% selecting test-data generation, 37.8% selecting BDD scripts, and 30.1% selecting load tests. (Tester and QA - IT Community Survey Report 2025)
Other categories were less frequently selected. Scripts for logging in and monitoring data were selected by 29.6%, unit tests by 20.9%, and security tests by 9.2%. An “other” category accounted for 4.6%, end-to-end tests for 2.0%, and no test scripts for 0.5%. These are selections among people who write automated tests, and the categories describe what respondents selected rather than a mutually exclusive portfolio. (Tester and QA - IT Community Survey Report 2025)
The 2018–2019 Sauce Labs/DZone results provide additional context for how automated portfolios changed. Automated component-test adoption rose from 52% in 2018 to 59% in 2019. Automated story-level-test adoption rose from 36% to 46% over the same period. Automated performance-test adoption moved in the opposite direction, falling from 51% to 46%. These figures track reported adoption across survey years; they do not establish why the changes occurred. (DZone Guide to Automated Testing: Sauce Labs Spotlight)
Integration testing in delivery pipelines
Integration tests are often discussed alongside test-driven development and continuous delivery, but the available figures measure related practices rather than proving that one causes another. In the Sauce Labs/DZone survey, 28% of respondents said they practiced TDD across all development efforts in 2018; that share was 22% in 2019. The share reporting no TDD declined from 36% to 34%, while the share using TDD only in certain teams increased from 36% to 45%. (DZone Guide to Automated Testing: Sauce Labs Spotlight)
Team structure was associated with different TDD reporting in the 2019 results. Among respondents using TDD on all projects, 15% also reported a dedicated DevOps team, compared with 7% among respondents without a DevOps team. The survey also recorded several reported benefits of TDD: 80% cited improved code quality, 59% cited less debugging time, 49% cited more maintainable code, and 45% cited using tests as documentation. These are respondent-reported benefits, not independently measured outcomes. (DZone Guide to Automated Testing: Sauce Labs Spotlight)
Pipeline structure was also reported in the same 2019 survey. Fifty-eight percent said their builds were broken into stages, a structure that can provide separate points for integration-test execution. Forty-six percent said their builds included automatic checks to proceed. The figures indicate the presence of pipeline practices; they do not specify whether every integration test ran in a dedicated stage or whether the checks applied to integration testing. (DZone Guide to Automated Testing: Sauce Labs Spotlight)
Who owns integration testing
Testing responsibility is distributed across development and quality roles in the State of Frontend 2022. Developers and QAs shared testing responsibility according to 43.4% of respondents. Only developers were responsible according to 24.5%, while 20.3% said mostly developers were responsible. Mostly QAs were responsible for 9.5%, and only QAs for 2.3%. Overall, 88% said developers were at least as involved in testing as QAs. (The State of Frontend 2022)
GitLab’s 2020 DevSecOps report measured a different organizational view. Among testers, 56% said QA teams were embedded with developers, 22% said QA was a separate organization, and 15% said QA was part of operations. The same report found 26% saying they had access to monitoring data because of their DevOps structure, 21% reporting visibility across development, security, and operations, and 18% reporting sufficient automation. Only 13% said their workflows did not get in the way. (2020 DevSecOps Report)
Staffing responses add a time-bound view of the operating model. In the 2020 report, 58% said QA staffing had stayed the same over the prior year, while 25% said it had increased. The report’s staffing figures are not specific to integration testing, but they help describe the organizational conditions in which integration-test work may be assigned and maintained. (2020 DevSecOps Report)
Why survey results vary
A literature review of industrial surveys found substantial variation in integration-test usage. Across its non-Likert primary-study results, the minimum was 37%, the maximum was 82%, the mean was 65%, and the median was 72%. The standard deviation was 19 percentage points. The review therefore found that integration-test popularity could be as low as 37% in one surveyed population and as high as 82% in another. (Industrial Surveys on Software Testing Practices: A Literature Review)
The review also summarized a 2021 study of GUI practitioners in Brazil. In that population, 28% reported automating integration tests and 37% reported automating API tests. These results are specific to the study population and should not be treated as a global estimate. The review further reported a 65% mean for integration-test usage versus 79% for unit-test usage across its non-Likert primary-study results. (Industrial Surveys on Software Testing Practices: A Literature Review)
Automation is a separate dimension from usage. Across the reviewed surveys, automation tools were used less than 50% of the time for each test level, including integration testing. Thus, a team may report using integration tests without reporting that all, or even most, of those tests are automated. The review’s figures cover primary studies reviewed through 2022 and combine results from different survey populations. (Industrial Surveys on Software Testing Practices: A Literature Review)
What the figures show
The strongest consistent signal is that integration testing is established but not uniform. Developer-focused surveys in 2022 reported 61.5% and 63%, while automated-testing users reported 77% and 78% in 2018 and 2019. The literature review’s broader range, from 37% to 82%, shows why a single benchmark can be misleading.
Portfolio data also separates frequency from balance. Integration tests were selected by 66.3% of automated-test writers in the 2025 IT Community survey, matching acceptance tests, while GitLab’s 2019 recorded inventory contained 17.9% integration tests in aggregate. Those measures have different denominators and scopes, so they describe different aspects of testing practice.
Finally, ownership and delivery context matter. Developers were at least as involved as QAs for 88% of State of Frontend respondents, and 56% of testers in GitLab’s 2020 report said QA was embedded with developers. At the pipeline level, 58% of Sauce Labs/DZone respondents reported staged builds and 46% reported automatic checks to proceed in 2019. Together, these figures describe an operating environment in which integration testing is common, but its automation, placement, and responsibility depend on the team and survey population.