Once in a while, I embark on an extensive research about the most fresh, up-to-date statistics around the application integration / system integration sector. Knowing what integration issues companies face nowadays helps better understand our potential customer pain points and needs, shape a better offer and align our product map with the most current requirements. And since such research takes me a LOT of hours, this time I decided – I can catch two birds with one hand (I know, I know, this is not the original saying but I don’t like even the idea of killing birds); do my research and write a blog article based on them.
So, here are some interesting data over the past 2-3 years and the conclusions we can derive from these. I grouped the insights in two parts: the ones that find support by data from several studies, and the ones derived from single studies.
Part 1: An increasing number of applications per company drive the need for application integration even further
Companies continue using a large number of applications
…and it doesn’t seem to be any different any time soon. Over the past several years, we’ve all seen reports saying that an average company uses XY number of applications, and if you can at least roughly recall the numbers, this tendency has only been growing.
A new joint survey by Mulesoft and Deloitte1 sent ripples all over by stating that companies now use an average of 976 applications, which is by 133 apps more than a year ago. It’s also worth mentioning that only 28% of these are integrated, according to this new report.
And I have to say, this is quite a shocking number. Maybe, it’s because of the size of the companies that were part of the survey – all respondents came from large enterprises with at least 1,000 employees –, or the specific sample of the respondents for this particular study, or the semantics of the original survey question, because the study actually mentions ‘976 discrete applications‘ without clarifying what exactly they meant here by discrete. But still, this number is stunning, leaving one almost in disbelief.
In contrast, the recent 2022 Businesses at Work report by Okta2 cites a much lower number, even for large organizations (over 2,000 employees) – 187 applications deployed on average. But even here, it is an increase compared to the previous 2021 report (which cited 175 applications on average), and a steady increase in the average number of used applications in general, regardless of the size of a company.
It is hard to say which number is closer to reality. There are some other scattered studies I was able to find that support Okta’s findings. For example, Beamy, a SaaS Management provider, conducted its own research, according to which in 2021, companies had an average of 190 enterprise SaaS application solutions in use. This number is cited in this article but I wasn’t able to find the original research data, hence the sample size and the company types remain unknown. And Blissfully’s 2019 report3 found that companies used an average of 137 SaaS apps, which already was a 30% increase from 2018, leading to believe that this number has only grown further since then.
On the other hand, a somewhat older report (2017) by Cloud Security Alliance4 revealed that large enterprises with over 50,000 employees run an average of 788 custom applications, that is – application systems and services that were developed in-house5. This older study would, on the contrary, support the data provided by the study from Deloitte and Mulesoft.
I guess the truth lies – as is often the case – somewhere in the middle. I can hardly imagine that an average large company deploys less than 200 application software systems; if so, then this number certainly doesn’t include shadow IT that flies under the radar. But the number of 976 raises some doubts, either.
This is also a good reminder that we need to be careful with semantics when we design surveys and studies – unless ambiguity is the goal. It might, for example, very well be that the shocking number of 976 applications given by Mulesoft and Deloitte is correct after all but it covers any type of application, both third-party, off-the-shelf business software solutions as well as homegrown, custom-developed applications and services, and maybe even the aforesaid shadow IT.
Application integration (aka system integration) is still one of the top challenges for IT
While it was hard to find any industry-neutral numbers, Ascend2’s survey6 revealed that 44% of marketers they asked named inadequate or missing application integrations as number one challenge for succeeding in generating valuable insights from data. In the HR sector, the number is nearly the same – 45% of a 2021 Jitterbit’s survey7 respondents cited accessing and combining data from several HR-related systems as a problem.
Yet another 2021 study – this time by Deloitte8 – showed that poor integration across applications is seen as the second main barrier to the effective application of digital technology in procurement, with cumulatively 45% of respondents ranking this factor as one of their top three issues. It was beaten only by ‘poor data quality’ that “won” the first placement in the chart, although we have to acknowledge here that poor data quality and poor application integration often go hand in hand.
If we can extrapolate from these isolated reports, then it will be fair to assume that roughly over 40% of companies see application integration as one of their top challenges. And in fact, data from the Mulesoft & Deloitte’s 2022 report would support this assumption. According to it, 38% of respondents see integration of siloed applications and data as their biggest challenge to digital transformation within their organization.
Part 2: Top integration scenarios, EiPaaS on the rise, and the high costs of poor integration
Top five application integration scenarios
For the lack of a fresher report, here are the findings from the Gartner’s 2019 Business Application Integration Survey9:
- Analytical or BI environments (cited by 86% of the survey respondents)
- Legacy systems (79%)
- Partners and ecosystems (68%)
- Creation of a single source of truth (65%)
- Human Capital Management i.e. HR-related applications (63%)
I doubt that the first scenario has since lost its placement. After all, data intelligence, business intelligence, and data-driven decisions are still very high on any company’s agenda.
But I’d be curious to know if the second scenario (legacy systems) and the third one (partners and ecosystems) swapped placements.
More and more companies are talking about creating API ecosystems; SaaS vendors are growing to appreciate the power of SaaS ecosystems. So, it might very well be that the scenario ‘partners and ecosystems’ has gained in urgency since 2019. It’s a pity Gartner doesn’t have a new, updated version of this survey yet, but maybe they will bring it out soon.
EiPaaS is the driving force behind enterprise process automation
According to yet another Gartner’s recent survey, the 2021 Hyperautomation Survey10, 11% of business technologists (BTs) see an enterprise integration platform (EiPaaS) as one of the three most-used tools to facilitate business-driven automation initiatives.
At the same time, the survey also found that these BTs on average employ the capabilities of two vendors’ solutions.
This implies that a single tool may not be sufficient and actually, correlates with what Gartner has been highlighting in its report for several years – no one tool on the current market is capable of meeting all application integration requirements.
This might change in the future, though, when we’ll see more iPaaS / EiPaaS market consolidation. The first roll of acquisitions has already swept through this industry, although currently, the tendency seems to be that it’s SaaS vendors that are more interested in strengthening the integration capabilities of their products and services than iPaaS vendors wanting to expand their portfolio and beef up their technology stacks.
What makes me think this way? Well, a larger portion of the most recent acquisitions have been carried out by SaaS vendors: UiPath (robotic process automation platform) acquired Cloud Elements (2021), Salesforce – Mulesoft (2018), Workiva (cloud based work process management platform) – OneCloud (2021), Qlick (data visualization and business intelligence platform) – Blendr.io (2020), Microsoft took over Clear Software (2021), HubSpot – PieSync (2019), Notion – Automate.io (2021). This is an interesting tendency that would support the growing importance of an embedded integration platform, that is maybe even worth exploring in a separate article.
Poor integration costs businesses money… a lot of it
One more recent study, this one by Cleo11, found that a quarter of supply chain companies may lose over $ 500K per year because of poor business process integration. Out of 323 companies who participated in this survey, 85% openly stated that they are indeed losing money due to integration-related issues. Moreover, those who said they lose over $500K a year, were 24% of respondents – an uptick by 6% compared to the survey from the year before.
In this context, it’s interesting to note that according to the same survey, the number of companies who feel it takes them too much time to onboard new supply chain partners has also increased by roughly around 10%.
While there are different reasons for that, poor application integration is certainly one of them. I’ve already touched on this topic in my comparison of EDI vs API technologies a few weeks ago, and to sum it up here: It’s often due to incompatibility of old, legacy systems or custom-built applications and lack of integration tools to speed up the process of data transformation and data exchange between different protocols and APIs.
Unfortunately, I wasn’t able to find any fresh reports or surveys from other sectors, but considering that analysts and experts often criticize inadequate application integration, for example, in Human Capital Management (HCM), or in ecommerce / retail, we can assume that these sectors experience very similar money loss as described in Cleo’s survey for supply chain.
All 12 application integration statistics summarized
Be warned: If you have jumped straight to this part, you’ll see only the “naked” numbers. If you’d like to have more background data, details and conclusions based on these numbers, start from the top of the page.
On applications usage:
- Deloitte and Mulesoft (2022): companies now use an average of 976 applications, with only 28% of these to be integrated
- Okta (2022): large organisations have 187 applications deployed on average
- Beamy (2021): companies have an average of 190 enterprise cloud based application software solutions in use
- Blissfully (2019): companies use an average of 137 business cloud based apps
Note: Scroll to the top to understand what to make of these contradicting stats.
On the importance of application integration:
- Ascend2 (2018): 44% of marketers see poor integration as top challenge for succeeding in generating valuable insights from data
- Jitterbit (2021): 45% HR specialists identify accessing and combining data from several HR-related business software systems as one of top problems
- Deloitte (2021): 45% respondents in procurement cite poor integration as the second main barrier to the effective application of digital technology
- Deloitte and Mulesoft (2022): 38% of respondents see integration of siloed business software applications as their biggest challenge to digital transformation
Other isolated statistics:
- In 2019, Gartner cited analytical or BI environments as the top application integration scenario (86% of respondents), followed by legacy software systems integration (79%), partners and ecosystems (68%), building of a single source of truth (65%), and process automation in Human Capital Management sector (63%)
- 2021 Hyperautomation Survey (Gartner) revealed that 11% of business technologists consider an EiPaaS as one of the top three tools they use most
- On average, business technologists use at least two vendors’ application solutions to drive business processes automation within their organizations
- Cleo’s 2022 report found that 85% of supply chain companies believe they’re losing a lot of money because of insufficient business software system integrations; out of these, 24% say they lose over $500K per year – nearly a quarter of respondents and an uptick by 6% from the year before
Conclusion: Application aka system integration may shape an enterprise’s survivability
At the risk of sounding overdramatic, I’d say that based on the most recent surveys and reports, system integration (or ‘application integration’ if you want) is likely to become the make-or-break type of thing for enterprises – if this is not the case already.
We can see that an average number of applications per company is only growing from year to year. We also see that enterprises use a large number of not only off-the-shelf business applications that are more likely to provide in-built integration capabilities, but also a huge number of homegrown applications that still need to be integrated into an enterprise’s data ecosystem.
And this leads to the growing importance of using an integration tool to let all these customer, product, partner and other data flow easily from one application to another. Personally I see the purpose and the need of such tool not only in facilitating the creation of integrations and automations per se. I would argue that its most value will be, in fact, in integrations maintenance and management; in providing IT teams with a holistic view of all the connected systems, integration projects and data flows organization-wide, and in enabling them to cooperate on existing projects or quickly get up to speed on new ones regardless of the geographical location and the level of involvement.
5. A side note: another number I found was “900 application software per midsize and large company on average”; the blog article that cited this number referred to some 2021 Talend platform and Cohesity Data Health survey, but the linked documents didn’t contain any data on this and it didn’t even look like these two companies conducted a joint survey. So, unless you’ll be more successful with your research and actually find the survey where these numbers originated from, I would discard it as a product of a content marketer’s imagination.
10. The original document of this particular study was nowhere to be found but at least, Gartner themselves refer to it in one of their reports: https://www.gartner.com/doc/reprints?id=1-27ZQY0UE&ct=211109&st=sb