Make’s 2023 year in review reported rapid community growth, more than 4 billion scenario runs, hundreds of apps added, and product releases intended to make scenarios easier to organize and reuse. For an operations team considering lead intake automation, the practical lesson is not to use those platform-wide figures as a forecast. Define the incoming data, consent checks, duplicate rule, destination, and exception owner for the specific workflow.
In its December 2023 retrospective, Make reported that its community grew 60% to nearly 2 million users and that it processed more than 4 billion scenario runs in 2023, nearly twice the 2022 total. It also said more than 300 apps were added during the year and that its Apps Marketplace featured more than 60 new apps. The retrospective does not provide counting methodologies or explain how those app figures relate. These are Make reported historical aggregates, not independently audited measures of successful business outcomes, unique users, or distinct scenarios.
What Make’s 2023 milestones do and do not tell you
The figures indicate the scale Make reported for its ecosystem in 2023. They do not establish how reliably a particular scenario will run, what it will cost, whether a specific app connection is available today, or whether an automation will improve a business metric. Make identified Google Sheets, Gmail, Airtable, Google Drive, Notion, Slack, Telegram, and Facebook Lead Ads among frequently used apps. It described OpenAI, Anthropic Claude, Jasper, and EdenAI as fast growing AI related apps, but supplied no underlying rankings, percentages, or usage methodology.
Platform wide adoption shows that automation is being used at scale. It does not prove that a new workflow will be reliable, accurate, or valuable.
What the 2023 product releases mean for workflow design
Make highlighted Custom Scenario Properties for organizing scenarios, Custom Functions for transformations, Scenario Inputs for supplying data to reusable scenarios, variables, and the ability to order router routes. Those announcements describe capabilities Make featured in 2023. For implementation decisions now, use the current documentation and confirm plan eligibility rather than assuming that historical terms still apply.
- Organize scenarios: Custom Scenario Properties let teams add metadata and filter or sort scenarios. Current documentation lists the feature for Enterprise customers. Metadata can identify an owner or environment, but it does not validate runtime data or enforce an approval process.
- Define reusable inputs: Scenario Inputs can be supplied manually, passed from another scenario, or sent in a Make API request body. The Start scenario module is documented for subscenarios. Required fields and defined data types provide structural checks, not proof that a business action is permitted.
- Route by explicit conditions: A router sends data through filtered routes in their configured order. Current documentation says routes run sequentially, not in parallel. Use filters for fixed rules and a fallback route for records that do not match known conditions. Route ordering is not a transaction or duplicate prevention mechanism.
- Transform bounded data: Current Custom Functions documentation lists Enterprise availability and JavaScript ES6 without third party libraries. Functions run synchronously and cannot make HTTP requests or use asynchronous code. They suit small, testable transformations, not general integrations.
- Store nonsecret configuration: Current custom variable documentation lists Pro, Teams, and Enterprise availability. Variables persist across runs, but values are stored in plain text. Do not put passwords, tokens, or other secrets in them. Use appropriate connection and credential mechanisms instead.
Make also said its AI Assistant had entered private alpha in December 2023. That historical announcement does not establish the assistant’s present availability or behavior.
Turn a marketing automation example into a controlled workflow
Make’s marketing examples article describes patterns such as emailing new Facebook Lead Ads leads, creating Salesforce leads from Eventbrite attendees, and tracking Instagram metrics in Google Sheets. These are illustrative editorial examples, not verified current templates or complete production builds. The article does not establish current trigger names, field mappings, permissions, consent checks, deduplication, or error handling.
The table below translates three examples into hypothetical workflow designs. Confirm exact app modules, access permissions, and source availability for the systems selected before implementation. A social metric workflow also requires an authorized source for the observation. The example alone does not establish a current Instagram connection.
| Workflow | AI responsibility | Validation and action | Failure owner |
|---|---|---|---|
| Lead follow up: new source lead event, identity and consent fields, existing CRM lookup, approved CRM update or follow up. | Optional classification of free text intent only. AI does not decide consent or authorization. | Require identity and consent data, check the duplicate key and permitted action, then save the source event ID and proposed or completed action with the CRM record. | Marketing operations reviews missing consent or conflicting identity data instead of allowing an automatic write. |
| Event attendee to CRM: attendee event, contact mapping, existing record check, CRM create or update, owner notification. | None for routine field mapping. | Check required identity fields and match against a stable source ID or other approved key. Save the event ID and CRM record ID for traceability. | Sales operations decides whether an ambiguous match should be merged, updated, or created. |
| Competitor metric tracking: scheduled or otherwise available authorized observation, account and timestamp validation, metric observation write. | None for numeric capture. | Define one row as one observed metric for one account at one timestamp. Store the account identifier, metric name, value, observation time, and source reference. | Operations investigates missing or implausible values and confirms the source and measurement period. |
For duplicate sensitive writes, a lookup followed by a create can still race. Two concurrent runs may both find no record and then create one. Where the destination supports it, use an idempotency mechanism or atomic upsert. Otherwise, use a database enforced unique key or transactional upsert where available. A Make router does not provide that database level guarantee.
If CRM ownership, record quality, or destination design needs broader attention, see CRM systems consulting.
A reusable scenario needs an input contract, not just a trigger
Scenario Inputs make it possible to define what a reusable scenario expects at its boundary. For example, a lead intake scenario might accept the following proposed fields. This is an illustrative contract, not a Make published schema:
{
"source_system": "lead_source",
"source_event_id": "evt_illustrative_1042",
"email": "[email protected]",
"consent_status": "granted",
"received_at": "2026-10-16T10:30:00Z"
}Define which fields are required, their data types, and, for collections, the expected structure. An unspecified collection structure can weaken structural validation. After input validation, apply separate business checks for consent, identity, authorization, and destination specific requirements. A correctly typed email value does not prove that the address exists or that a downstream CRM will accept the write.
Structural validation answers whether data has the expected shape. Business validation answers whether the value is allowed and the action is authorized. Keep those checks separate so a technically valid input cannot bypass a policy gate.
Customer outcomes and community stories are examples, not benchmarks
Make’s 2023 retrospective attributed several results to EdTech company Tutero: one person managing more than 250 tutors, more than 20 hours of finance work saved per week, a 50% reduction in failed payment issues, and average lead response times of 25 minutes or less. These are Make reported customer story claims, not independently verified measurements. The supplied Tutero story link now redirects to Make’s general Success Stories index, and no standalone current story was verified.
The transferable lesson is that moving information between systems can reduce repetitive work, but results depend on the starting process, data quality, ownership of exceptions, and measurement method. Before rollout, record a baseline, define the same metric and observation period for review, and decide who investigates failed or ambiguous records. Make’s mention of frequently used apps and fast growing AI related apps, including OpenAI, is similarly a company description without published rankings or usage methodology.
Integromat’s transition and Waves ’23
Make’s official Integromat notice said support would end on June 30, 2023, and that the platform would shut down on September 30, 2023. It described migration to Make and said scenarios would be deactivated at shutdown. The notice does not establish that every remaining user completed migration.
For any legacy workflow still being reviewed, verify the live scenario, credentials, trigger endpoint, and downstream effects in the current environment. A safe review should also test webhooks, mailhooks, OAuth connections, schedules, filters, error handling, and downstream writes. Make described Waves ’23 as its first in person user event, held in Munich with hundreds of attendees and nearly 30 speakers and hosts. Those details come from the company retrospective.
Practical takeaway: design the process before choosing automation
Make’s reported 2023 scale and product breadth illustrate what a broad automation platform can support. They do not replace process ownership, validation, or exception handling. Before building, decide which event is authoritative, what counts as a duplicate, which system owns the record, what permissions allow the write, and who resolves exceptions.
- Name the source event and system of record.
- Define required data, consent checks, and the duplicate key.
- Confirm write permissions and an idempotent or uniqueness strategy where needed.
- Assign a person to own the exception queue.
- Choose a baseline and review period for the intended outcome.
If those decisions are clear and the required app modules and permissions are confirmed, proceed with representative tests, including missing fields, duplicate events, ambiguous matches, and rejected writes. For support designing or reviewing a Make workflow, explore Make automation consulting.
