Half of B2B website tools are gone within 4 years
Half of the tools B2B software companies add to their homepage are gone within 4 years. Live chat peaked on 14.5% of homepages in 2019 and is on 7.8% now.
Methodology (PDF)Key findings
44.3%
of 18,951 new tool installs on B2B software homepages were still live four years later. 79.3% lasted to the next yearly snapshot.
-46%
for live chat, from 14.5% of homepages at its 2019 peak to 7.8% in 2026. Drift fell from 2.6% in 2020 to 0.2%.
8 years
median life of a new tag manager install, against 3 years for chat, cookie consent, forms, marketing automation and scheduling tools.
63.5%
of homepages ran a tag manager in 2026, up from 3.4% in 2014. Cookie consent banners grew from 0.3% to 16.8%.
On this page
Half of the website tools B2B software companies add to their homepage are gone within 4 years. We followed a random sample of 10,000 B2B software companies through 55,285 yearly homepage snapshots from 2014 to 2026 and recorded which marketing and analytics tools were live on each page. Of 18,951 new tool installs, 79.3% were still there at the next yearly snapshot, 53.0% after three years and 44.3% after four.
Share of 18,951 new tool installs on B2B software homepages still live after each number of years (Kaplan-Meier estimate). Installs are timed from yearly snapshots, so each duration is an upper bound by up to a year. The line at 50% marks the median.
Source: Kenso internal data (Kenso's archive of company websites and the company records Kenso uses for sales and outreach), analysed by Cristian Frunze.
Show the dataHide the data
| Years since the tool appeared on the homepage | New installs still live |
|---|---|
| 0 | 100.0% |
| 1 | 79.3% |
| 2 | 64.3% |
| 3 | 53.0% |
| 4 | 44.3% |
| 5 | 37.2% |
| 6 | 31.8% |
| 7 | 27.1% |
| 8 | 22.4% |
| 9 | 19.2% |
| 10 | 16.0% |
What we measured
The companies are software, internet, security and data firms with 11 to 5,000 employees, headquartered in the US, the UK, Canada, Australia, Ireland or New Zealand, drawn from the company records Kenso uses for sales and outreach. For each one we took one homepage snapshot a year from Kenso's archive of company websites, and kept pages that loaded, were not parked and had at least 50 words of text.
Kenso's technology detector read each page. A tool counted only when its loading script was live: a tag left in an HTML comment, a plain link or a mention in the text did not count. We tracked seven categories: analytics, tag managers, marketing automation, forms and surveys, live chat, scheduling and cookie consent. A new install is a tool that appears after a company's first usable snapshot. It ends at the first later snapshot without it.
Live chat boomed, then lost almost half its homepages
The clearest boom and bust is live chat. In 2014, 6.0% of these homepages ran a chat widget. By 2019 it was 14.5%. It has fallen every year since, to 7.8% in 2026: 46% below the peak.
- Any live chat
- Intercom
- Drift
- Any live chat
- Intercom
- Drift
Share of usable B2B software homepages with a live chat widget loaded, by year, and two of the best-known chat tools, Intercom and Drift. A tool counts only when its loading script is live on the page.
Source: Kenso internal data (Kenso's archive of company websites and the company records Kenso uses for sales and outreach), analysed by Cristian Frunze.
Show the dataHide the data
| Year | Any live chat | Intercom | Drift |
|---|---|---|---|
| 2014 | 6.0% | 0.0% | 0.0% |
| 2015 | 7.1% | 0.3% | 0.0% |
| 2016 | 8.7% | 1.5% | 0.1% |
| 2017 | 10.5% | 2.8% | 0.8% |
| 2018 | 12.9% | 4.0% | 1.8% |
| 2019 | 14.5% | 4.7% | 2.5% |
| 2020 | 14.1% | 3.7% | 2.6% |
| 2021 | 12.7% | 3.2% | 2.0% |
| 2022 | 11.9% | 2.7% | 1.5% |
| 2023 | 10.6% | 2.4% | 1.2% |
| 2024 | 9.5% | 2.5% | 0.7% |
| 2025 | 8.7% | 2.2% | 0.4% |
| 2026 | 7.8% | 2.2% | 0.2% |
Drift, which popularized "conversational marketing", went from no homepages in our sample in 2014 to 2.6% in 2020, then fell to 0.2% in 2026. Intercom peaked at 4.7% in 2019 and is on 2.2% now.
Companies did not swap one chat widget for another. When a chat tool came off a homepage, a different chat tool went on in the same year only 12.0% of the time. Most of the time, the homepage simply stopped offering chat.
One objection: maybe the widgets moved inside a tag manager, where the page code does not show them. On company-years with no tag manager at all, chat fell even harder, 63% from its peak.
Tag managers and consent banners kept spreading
Two kinds of tool spread through the whole period. Tag managers went from 3.4% of homepages in 2014 to 63.5% in 2026. Cookie consent banners went from 0.3% to 16.8%, with the jump in 2018, the year Europe's GDPR took effect: from 1.0% in 2017 to 4.5%. Since 2019 they grew 2.47 times, the fastest relative growth of any category.
- Tag managers
- Cookie consent
- Any analytics
- Tag managers
- Cookie consent
- Any analytics
Share of usable B2B software homepages with at least one tool in each category, by year. Analytics loaded through a tag manager does not show in the page code, so part of the analytics decline may be tools moving inside one.
Source: Kenso internal data (Kenso's archive of company websites and the company records Kenso uses for sales and outreach), analysed by Cristian Frunze.
Show the dataHide the data
| Year | Tag managers | Cookie consent | Any analytics |
|---|---|---|---|
| 2014 | 3.4% | 0.3% | 71.9% |
| 2015 | 6.5% | 0.4% | 73.1% |
| 2016 | 11.1% | 0.8% | 72.9% |
| 2017 | 13.0% | 1.0% | 69.7% |
| 2018 | 27.0% | 4.5% | 67.7% |
| 2019 | 40.4% | 6.8% | 66.7% |
| 2020 | 46.6% | 8.0% | 62.2% |
| 2021 | 55.1% | 8.8% | 59.8% |
| 2022 | 60.4% | 10.5% | 59.9% |
| 2023 | 63.8% | 12.1% | 59.0% |
| 2024 | 65.4% | 14.1% | 57.9% |
| 2025 | 65.0% | 15.0% | 56.6% |
| 2026 | 63.5% | 16.8% | 55.9% |
Marketing automation roughly doubled, from 10.1% to 21.7%, almost all of it HubSpot, which grew from 3.8% to 16.7% of homepages.
The biggest single change was forced. Google retired Universal Analytics in 2023 and moved everyone to Google Analytics 4. Universal Analytics was on 73.6% of these homepages at its peak and is on 12.6% now, while GA4 reached 29.7%. Many companies that dropped the old tag show no new one in their homepage code. Any analytics tool at all fell from 71.9% of homepages in 2014 to 55.9% in 2026.
Infrastructure lasts, features churn
Across 40 widely used tools, the median tool lost 28.4% of the homepages it was on each year. The infrastructure tools are the exception: Google Tag Manager lost 9.0% a year and Google Analytics 11.7%.
Median years until half of the new installs in each category were gone (Kaplan-Meier). Scheduling rests on the fewest installs, 151; analytics on the most, 5,844.
Source: Kenso internal data (Kenso's archive of company websites and the company records Kenso uses for sales and outreach), analysed by Cristian Frunze.
Show the dataHide the data
| Median years until half are gone | |
|---|---|
| Tag managers | 8 |
| Analytics | 4 |
| Cookie consent | 3 |
| Forms and surveys | 3 |
| Live chat | 3 |
| Marketing automation | 3 |
| Scheduling | 3 |
Half of new installs were gone within 3 years for chat, cookie consent, forms, marketing automation and scheduling tools, within 4 years for analytics and within 8 years for tag managers.
What we expected
We wrote down five predictions before reading any snapshot.
- Tools fall faster than they rise. Wrong. Of 7 tools with a measurable rise and fall, 2 declined faster than they rose. Live chat took 3.9 years to climb from half its peak to the peak and, 7 years later, is still above half of it.
- The median install lasts under 4 years. Not as stated: the median is exactly 4. Chat and marketing automation (3 years) do churn faster than analytics (4) and tag managers (8), as expected.
- Consent tools spread fastest from 2019, more than doubling. Right on relative growth (2.47 times). Tag managers gained more percentage points (23.1 against 10.0).
- Live chat peaked and fell at least 20%. Right: 46% below the peak.
- Replacements go to the big suites. Mostly wrong. HubSpot, Google and Salesforce took 52.6% of replacement installs in marketing automation and 39.2% in forms, but 20.1% in analytics and almost none in chat or consent. The largest single flow was Contact Form 7 to HubSpot Forms, at 81 companies.
What this data cannot show
- Homepage code only. Tools inside a product, on other pages or loaded by a tag manager are not seen. Some declines could be tools moving inside a tag manager, though the chat decline holds where none is present.
- Installed, not used. A live loader shows a tool was installed, not that anyone used it or paid for it.
- Survivors only. The sample is companies that exist today, so companies that closed earlier are missing from the early years.
- Yearly snapshots. Install and removal dates are known to the year, so durations are upper bounds. The true median can only be shorter.
- Uneven coverage. The archive holds fewer snapshots for early years: 2,570 usable homepages in 2014 against 5,103 in 2022. Industry labels also do not separate B2B from B2C software.
What it means for B2B teams
- Budget the exit when you buy. Half of new homepage tools are gone within 4 years, and most categories turn over in 3. Treat a website tool as a trial with a renewal date, not a platform decision.
- A category can shrink while its vendors grow. Live chat has lost share on B2B software homepages every year since 2019. If you sell chat, you are selling into a shrinking base.
- Infrastructure sticks. Tag managers last 8 years at the median, more than twice as long as the tools they load. Switching costs live in the layer underneath.
- Refresh tech-stack data often. A typical tool loses 28.4% of its homepages each year, so a list of who runs what goes stale fast.
Data and method
- Data
- Kenso internal data: Kenso's archive of company websites, one homepage snapshot per company per year, and the company records Kenso uses for sales and outreach, reused by Cristian Frunze for this study. Companies headquartered in the US, the UK, Canada, Australia, Ireland and New Zealand.
- Sample
- 55,285 usable yearly homepage snapshots from a random sample of 10,000 B2B software companies with 11 to 5,000 employees; 18,951 new tool installs across 40 widely used tools.
- Date range
- One homepage snapshot per company per year, 2014 to 2026.
- Method
- Tools detected with Kenso's technology detector and counted only when the loading script was live on the page. Yearly adoption shares, Kaplan-Meier survival of new installs and same-year switching. Headline numbers recomputed by an independent analysis.
- Excluded
- Parked pages and pages with fewer than 50 words of visible text
- Tools found only in HTML comments, plain links or page text
- Tools inside the product, on other pages or loaded only through a tag manager
- Any count under 50 companies
- Privacy
- Aggregates only. Any group smaller than 50 people is suppressed, and no individual is named.