Your software bill is going up, and it is probably not because you bought more project management. In Q2 2026, Vertice found that artificial intelligence went from 4.0% to 11.3% of small and medium business software budget in a single quarter, overtaking both security at 9.2% and development at 9.7% to become the largest category outright. That money came from somewhere. The same data has SMBs cutting budget share for security by 19%, marketing by 16%, and CRM by 11% to fund it.
So the consolidation conversation is live again, for a reason that has nothing to do with tidiness. There is a new line item that keeps growing, the total is not growing as fast, and something has to give.
Most advice on the subject is really just a target number. Get from thirty tools down to twelve. The trouble is that nobody agrees what the number is. Cledara’s SaaS Spend Index, built from real anonymized company payments rather than surveys, puts the typical company at 32 paid tools. At the other end of the scale, CIO Dive reported on Torii’s 2026 benchmark, which found the average large enterprise running 2,191 applications, with more than 61% of them outside any formal IT oversight and the average employee touching 40 to get their job done. Different populations, different counting methods, wildly different answers. Benchmarking yourself against any of them tells you nothing useful about your business.
The tool count is the wrong number to optimize
Two tools that never exchange data cost you two subscriptions. That is the entire cost. Merging them saves one subscription and buys nothing else, and if the combined product is weaker at either job, you have paid money to make your week harder.
Two tools that exchange data every day cost you two subscriptions plus a tax charged on every exchange. Somebody retypes a customer name. Somebody pastes half a conversation into a ticket so the next person has context. Somebody spends four minutes working out which of the two systems is telling the truth. None of that shows up on an invoice, which is exactly why it survives for years.
The return on consolidation comes from removing handoffs, not from removing logins. So the number worth counting is not how many tools you own. It is how many times a day work crosses a boundary between them.
Count your handoffs before you shortlist anything
This takes an afternoon and it will change your shortlist. Take the last twenty pieces of work that involved more than one tool and write down where the boundary was. You are looking for three distinct kinds of handoff, because they do not cost the same.
- Re-entry: the same information typed into two places. Cheap each time, expensive in aggregate, and every copy is a fresh opportunity to be wrong.
- Lookup: answering one question requires opening two tools. “Has this customer raised this before, and did we ever actually fix it?” is the classic, and it is the reason people give up and guess.
- Reconciliation: the two tools disagree and somebody has to decide which one is right. The most expensive of the three, because it costs judgement rather than minutes, and it quietly erodes confidence in both systems.
Score each one by how often it happens and how bad it is when it goes wrong. Nearly every small business ends up with two or three joins that genuinely matter and a long tail that does not. That short list is your consolidation project. The rest is noise, and treating it as scope is how these projects turn into a year of migration work with nothing to show.
Four questions to ask about each tool
Run every tool in your stack through these. Most will fall out at the first question, which is the point.
- Does anything in here need to reference a record somewhere else? If nothing in your design tool ever needs to point at a support ticket, leave your design tool alone. It is not part of this conversation.
- Does the connection need to be live, or would a monthly export do? A live link is one you follow while you are working, mid task. If you only need two datasets side by side once a month to produce a report, an export is a perfectly respectable integration and it costs nothing to build.
- Who owns this record type? Consolidation goes wrong when two systems both believe they own customers. Name one system of record per record type first, and be willing to make the loser read-only or switch it off. Skipping this is how you end up with three versions of a contact list and no way to tell which is current.
- What quietly breaks if you turn it off on a Friday? The forgotten inbox rule. The automation somebody built two years ago and never mentioned. The invoice that arrives from it. The export your accountant expects in January. Write these down before you migrate rather than discovering them during.
What consolidation should actually buy you
When you get to evaluating platforms, do not ask how many modules a product has. Ask what it lets you connect, then test your own two or three joins against it. Here is what that looks like in practice using ISO Mate, which is built as one dataset across support, work, sales, and compliance rather than as separate products behind one bill.
Requests that turn into work. A support ticket can be linked to issues, tasks, incidents, and emails, and the issue link works from both ends: the developer who opens the issue can see which support requests it addresses, and the agent can see the development work behind the answer they are about to give. For most small businesses this is one of the two or three joins that matter, and it is usually the one that is missing entirely.
A person’s whole history on one screen. Contacts are extracted automatically from email, so the directory builds itself. Open one and you get the messages where that person was sender or recipient across your shared mailboxes, alongside the help desk tickets where they are the requester, and you can start a new ticket with them already set as requester. It works from the other direction too: the email activity panel shows the related contacts, tickets, and issues for a message you are reading, and links it to a ticket or creates one from it.
Sales records with the follow-up attached. Tasks, calendar events, notes, and emails link to any lead, opportunity, or deal. The activity timeline records stage changes, assignment changes, linked notes, created tasks, calendar events, emails, and contact associations as they happen, without anyone maintaining it. This is the join most CRMs solve for themselves and nobody else, which is why the follow-up task usually ends up in a different tool from the deal it belongs to.
One screen for the morning. The dashboard is widget based, with 14 widgets covering every area, drag to reorder, resize between half and full width, seven role-based starting templates, and optional auto-refresh. Quick Actions at the top create an issue, task, test cycle, incident, ticket, or note without navigating anywhere. This is the part people underestimate: the daily cost of a split stack is mostly the twenty minutes of checking six places before you know what today looks like.
Automation that is not an integration project. This is the real payoff, and it only exists because the records already sit together. Workflow Automation reads and writes across more than 30 entity types with more than 30 action types, so a rule can start on a ticket and end on a task, a calendar event, a note, or an email. There are eight trigger sources: a record being created, updated, or deleted, a specific field changing value, a cron schedule, an offset before or after a date field, a record going untouched for a set period, a manual run from a record’s detail page, an inbound webhook, and a public form submission. If you have ever tried to build the equivalent across two vendors with a middleware subscription in between, the difference in effort is the argument for consolidating in the first place.
You do not have to design any of that from scratch. The template gallery includes ready-made workflows such as Overdue Task Reminder, Weekly Stale Tickets Digest, New Issue to Sprint, and Auto-Acknowledge Inbound Email, each instantiating with its full action tree intact. New workflows arrive disabled so you can read them before they touch anything, and dry run executes one against a real or synthetic record with no side effects at all.
What to leave exactly where it is
Some things should not be pulled into a work platform, and any vendor who tells you otherwise is selling rather than advising.
- Anything to do with money. Accounting, invoicing, payroll, banking, tax. Specialised, regulated, and already wired into how your accountant works. ISO Mate does not do these, and you should not want a project platform to.
- Deep specialist tools. Your code hosting, your design tool, your data warehouse. Trading a great specialist tool for an adequate module is a downgrade wearing the costume of a saving.
- Whatever the team is currently trialling with AI. Cledara has AI platform spend up 353% year on year among UK companies, the fastest rising category by a wide margin and several times the next one. That is an experiment in progress, and experiments do not belong in a consolidation plan until they settle down into something you would sign a year for.
Naming these matters, because consolidation is not all or nothing and it is usually sold as though it were. For the tools you keep, a webhook trigger lets an outside system start a workflow, and an HTTP request action lets a workflow call out to one with your own headers and body. You can join two systems without merging them, which is frequently the right answer.
The mistakes that make this cost more than it saves
- Migrating history nobody reads. Move open work and the most recent closed cycle. Keep a flat export of the rest and get on with your life. Full-fidelity history migration is where these projects go to die.
- Doing it all in one weekend. Move one join, live with it, then move the next. A stack half migrated on purpose is fine. A stack half migrated by accident because you ran out of weekend is not.
- Choosing on the feature grid. A long module list is not the same thing as the specific connection you need, and a grid will never tell you the difference. Build your two or three real joins during a trial, with your own data shapes.
- Rebuilding your worst process faithfully. A migration is the only sanctioned opportunity you will get to delete a step. Take it.
Questions worth putting to any vendor
- Can a record of type A reference a record of type B, and is the link visible from both sides? One-directional links just relocate the lookup problem to whoever is standing on the other end.
- Can an automation read in one area and write in another, and can I test it without side effects? The second half is the one people forget to ask, and it is the difference between trusting your automations and being slightly afraid of them.
- Can I report across areas, or only within one? Listen carefully to this answer. In ISO Mate, cross-area visibility lives on the dashboard, where widgets from every module sit on one screen, while the report dashboards are per module: DevOps, QA, tasks, help desk, incidents, compliance, sales, feedback, calendar, and account, each exportable to PDF or CSV. Knowing which is which before you buy saves a disappointing conversation in month three.
- Who can see what? Consolidating means your support contractor and your bookkeeper are now in the same system as everything else. Role-based permissions per area are what make that acceptable rather than alarming.
- How do I get my data out? Ask before you go in, not when you are leaving. In ISO Mate, lists and reports export to CSV and PDF with your current filters applied.
A sequence that tends to work
Pick the single handoff from your list that costs you most. Move both sides of that one join and nothing else. Run it for a full cycle, whatever a cycle means in your business, a month or a quarter. Then count that handoff again.
If it went to zero and nothing else got worse, you have your answer and you can move the next one. If it did not, you have learned something important cheaply, while you still have both systems and no sunk migration behind you. Either outcome is worth more than a plan to reduce your tool count by 40% that nobody will finish.
The honest version of this is that consolidation is not a tidiness exercise or a cost-cutting exercise. It is a decision about which records in your business need to know about each other. Answer that first and the shortlist writes itself.
If the join you care about runs between the requests you receive, the work they create, and the people who raised them, that is the seam ISO Mate is built along. Every account starts with a 14-day free trial, which is enough time to move one handoff and see whether the rest is worth doing.