How We Tested Jobber: Test Results and Evidence

STS-TEST-JOBBER-CORE-2026-001 first-use pass, September 1, 2026, completed as specified; observations confirmed by the driver.

• Reaching the New Client form took about 48 seconds on first use, most of it scanning the collapsed icon rail, which hides section names until hover, for the Clients section.
• The customer form is a single page with one save at the end; no intermediate confirmation appears before the record is created.
• From the task marker to the confirmed row in the client list: 341 seconds, including the form entry. The saved record and the list row showed the fixture values.
• No error, modal or upgrade prompt appeared during the four steps.

Capabilities exercised: crm.

STS-TEST-JOBBER-CORE-2026-002 repeat pass, September 2, 2026, not completed as specified.

• With the path known, the New Client form opened about 11 seconds after the task marker; the first-use cost of the collapsed rail disappears once the Clients location is learned.
• Protocol deviation: improvised customer data was entered instead of the fixture, so the success criterion (record matches the fixture) was not met and the run does not count as the repeat pass. The workflow itself completed without a product error.

This run is not admissible as a benchmark result: the fixture was not completed as specified (the run table has the row). Its observations above stand as recorded; it contributes no timing and no completed scenario.

STS-TEST-JOBBER-CORE-2026-003 repeat pass, September 2, 2026, completed as specified; observations confirmed by the driver.

• With the path known, the New Client form opened about 16 seconds after the task marker (dashboard, Clients, New Client).
• Saving a client that resembles an existing one raised a "Possible duplicate client" confirmation offering Cancel or Create Client Anyway; it neither blocked the save nor merged the records, and it took about 4 seconds to resolve.
• The save completed on the first try after the confirmation; the record and the list row matched the fixture. About 2 minutes 12 seconds from the marker to the confirmed row, including the confirmation.

Capabilities exercised: crm.

STS-TEST-JOBBER-CORE-2026-004 first-use pass, September 2, 2026, completed as specified; observations confirmed by the driver.

• The service catalog is not in the left navigation: it sits under the account menu, then Settings, then Products & Services. Reaching the item form took 306 seconds on first use, including a 27-second web search that was logged as help.
• The line-item pickers on the job, request and quote forms suggest catalog items and accept a typed name without creating one; they absorbed most of the search before the Settings path was found.
• The item form offers Item type, Name, Description, Cost, Markup, Unit Price, an image, a tax exemption and a booking duration; it has no category or tag field. The price was typed into Cost first and re-entered as Unit Price, because the three-field money row does not say which field is the customer price.
• Entry took 56 seconds; one Create; the confirmation and the list row appeared within about 3 seconds, and the reopened item showed the fixture values.
• Three "Discard unsaved changes" confirmations appeared, one per abandoned exploration form; no upgrade prompt or paywall touched the task.

Limitations observed

• A category or tag field on the Add New Product/Service form: the fixture's category "sweep" had nowhere to go (the form offers Item type Service / Product only); skipped per protocol

Capabilities exercised: estimating, online booking.

STS-TEST-JOBBER-CORE-2026-005 repeat pass, September 2, 2026, not completed as specified.

• With the path known, the item form opened 24 seconds after the task marker and the fixture was entered exactly in 22 seconds, the price landing in Unit Price on the first pass.
• Jobber refused the save with an in-form banner, "A product or service already exists with that name", greyed the Create button while the banner showed, and offered no way to proceed; the client form, by contrast, asks for confirmation on a possible duplicate.
• The collision was with our own earlier test item. The rename that got the save through put the created record off the fixture name, so the run does not count as the repeat pass.

This run is not admissible as a benchmark result: the fixture was not completed as specified (the run table has the row). Its observations above stand as recorded; it contributes no timing and no completed scenario.

STS-TEST-JOBBER-CORE-2026-006 repeat pass, September 3, 2026, completed as specified; observations confirmed by the driver.

• On a catalog cleared of the earlier test items, the item form opened 4 seconds after the task marker; the session was already on the Products & Services page, so this pass records no navigation time from the dashboard.
• The fixture was entered exactly in about 42 seconds; one Create; the "Service created" confirmation appeared within about 2 seconds and the list row within about 4; the edit view confirmed every value.
• The refusal seen in run 005 did not recur once the earlier item was deleted: it was a collision in our environment, not a form defect.
• The form still has no category or tag field, so the fixture category was skipped as the protocol allows.

Limitations observed

• A category or tag field on the Add New Product/Service form: the fixture's category "sweep" had nowhere to go (Item type Service / Product only); skipped per protocol, as in run 004

Capabilities exercised: estimating, online booking.

STS-TEST-JOBBER-CORE-2026-007 first-use pass, September 4, 2026, completed as specified; observations confirmed by the driver.

• From the dashboard, the path to a new quote took 15 seconds on first use: the Quotes link in the navigation, then the "Create Quote" card on the empty Quotes page. No help was needed.
• Picking the catalog item filled the line completely: name, description, quantity 1 and the unit price, with the item's non-taxable setting carried over, so the quote total matched the fixture before and after saving with nothing edited.
• Saving opened a "Send a test email" prompt on top of the new quote, with one "Review Test Email" button and a close control; it was closed without sending, and the quote stayed a draft. Whether it appears on every first quote is not yet known.
• The save control is a split button (a plain save and a dropdown); the plain save was used. The quote's first field is a free-text title.
• A first attempt 37 minutes earlier stalled on the Quotes page load because the connection was slow at the time; it was abandoned and the retry loaded the page in about a second.

Capabilities exercised: estimating.

STS-TEST-JOBBER-CORE-2026-008 repeat pass, September 4, 2026, completed as specified; observations confirmed by the driver.

• On the repeat, the path from the dashboard to a new quote took 14 seconds: the Quotes link, the "New Quote" button, then a templates prompt that had to be passed with "Create New Quote" (it did not appear in the first-use run, whose empty Quotes page used cards instead).
• Picking the catalog item filled the line completely again (name, description, quantity 1, unit price, the non-taxable setting), and the quote total matched the fixture before and after saving with nothing edited; the saved quote was on screen 56 seconds after the start marker.
• The "Send a test email" prompt seen after the first quote's save did not appear after this one; the save went straight to the saved quote. It reads as a first-quote prompt, not a per-save one.
• A second draft quote for the same client was accepted without a duplicate warning; the quote list showed both drafts at the fixture total.
• Every measured span was shorter than in the first-use run: form 15 to 14 seconds, line item 67 to 42, save 93 to 56, confirmation 128 to 60.

Capabilities exercised: estimating.

STS-TEST-JOBBER-CORE-2026-009 first-use pass, September 4, 2026, completed as specified; observations confirmed by the driver.

• From an existing quote, the path to a job on first use was the quote's More menu, where "Convert to Job" is the first item; the pre-filled job form was on screen 23 seconds after the start marker and the saved job at 59 seconds, with nothing typed.
• Jobber did not require an approval before converting: "Mark as Approved" and "Convert to Job" sit in the same menu, a draft quote converted directly, and the quote then read "Converted" with a link to the job rather than an approved state.
• The conversion's defaults set the visit to today, assigned to the account owner, and billing to "upon job completion", with an invoice row already waiting for the quote's amount; the job page's primary button offers to email a booking confirmation, which we did not use.
• One onboarding nudge appeared on the new job page and faded on its own; no other prompt interrupted the path.
• The quote's converted status was confirmed on screen after the timed session rather than within it, so the repeat pass ends on the quote page.

Capabilities exercised: estimating.

STS-TEST-JOBBER-CORE-2026-010 repeat pass, September 4, 2026, completed as specified; observations confirmed by the driver.

• Marking a quote approved from inside the account is one menu click with no confirmation dialog: the status changed at once, two brief notices faded on their own, and nothing was sent to the customer.
• Once a quote is approved, Jobber replaces the quote's primary "Send Email" button with "Convert to Job" and adds an approval date to the quote; the approved state showed immediately on the dashboard and in the quote list's status filter.
• Converting the approved quote was one click on that button: the job form opened pre-filled 21 seconds after the marker and the saved job 43 seconds after it, with nothing typed; the job carries the quote's number, the quote left the approved state, and the job page's primary button offers to email a booking confirmation, which we did not use.
• This repeat pass was completed in two recordings about twenty minutes apart on the same quote, the approval in the first and the conversion in the second, so no single workflow time is reported; the step timings stand on their own.

Capabilities exercised: estimating.

STS-TEST-JOBBER-CORE-2026-025 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• We created a new job directly from the jobs list rather than from an existing estimate, and the product did not offer a way to link the new job back to a quote on this path.
• The client picker showed each contact's name, address and status inline, which distinguished the right customer among near-duplicate names in the account.
• Adding the service line offered a shortcut to reuse the line item from the account's most recent job, alongside the normal catalog search.
• Saving the job took the browser straight to the new job's record, with the visit's technician field already defaulted to the account owner.

Capabilities exercised: scheduling.

STS-TEST-JOBBER-CORE-2026-026 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Starting this repeat run from an already-open jobs list, we reached the new-job form in a few seconds with no extra navigation.
• As before, the customer picker showed each contact's name and address inline, so the right customer was selected on the first pick despite near-duplicate names in the account.
• We confirmed the new job from the customer's own record this time, which lists all of that customer's jobs, quotes and invoices together, rather than from the general jobs list.
• The workflow again produced no confirmation message on save; the browser moved straight from the form to the finished job record.

Capabilities exercised: scheduling.

STS-TEST-JOBBER-CORE-2026-027 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Setting a job's initial visit time required unchecking an "Anytime" toggle before start and end time fields appeared alongside the visit's date.
• Finding the date/time editor took some exploration on first use: the visit's edit controls include a similarly placed arrival-window setting and a checklist setting next to the actual scheduling control.
• Changing an existing visit's date to a different day did not automatically update its time; the time fields had to be edited separately even when the date changed.
• Rescheduling the visit and confirming the new date and time took about four minutes from start to finish, including the time spent finding the right control on first use.
• No option to notify the client of the schedule change appeared on the editing screen; only an internal reminder setting for the assigned team member was offered.

Capabilities exercised: scheduling.

STS-TEST-JOBBER-CORE-2026-028 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Setting the visit's initial time required unchecking an "Anytime" toggle before start and end time fields appeared alongside the visit's date.
• On a known path, scheduling the visit took well under a minute from opening the job to saving the initial date and time.
• Changing the visit's date to a different day did not automatically update its time; the time fields had to be edited separately even when the date changed.
• The full schedule-then-reschedule workflow, including confirming the new date and time on the job's page, took about two minutes end to end.
• No option to notify the client of the schedule change appeared on the editing screen; only an internal reminder setting for the assigned team member was offered.

Capabilities exercised: scheduling.

STS-TEST-JOBBER-CORE-2026-029 first-use pass, September 9, 2026, completed as specified, with a recorded workaround; observations confirmed by the driver.

• In our standardized test, the account's team settings page showed a single assigned seat, already filled by the owner, and we did not open the seat-management or invite controls, so whether a second seat needs a paid plan on the trial was not tested.
• With no second user in the account, we assigned the account owner to the job's visit instead, to test the assignment workflow itself; the assignment showed up immediately in the visit's own confirmation view.
• Assigning a team member to a visit took a few clicks past the job page (its scheduled-visit list, an edit icon, then a save), with no separate confirmation step; the change was checkable right away in the visit record.
• Saving the same assignment a second time without changing anything produced a small notice that nothing had changed, rather than a generic success message.

Limitations observed

• A second (technician) user seat: the trial's Manage Team page (`/manage_team`) read "1/1 seat assigned," listing only the owner ("Assigned seats (1)": Evan Mercer, Owner, Active); the fixture technician "STS-TEST Tech One" was not added as a second user; the "Invite New Member" and "Manage Seats" controls were not opened, so whether a second seat needs a paid plan on the trial was not tested

Capabilities exercised: dispatch.

STS-TEST-JOBBER-CORE-2026-030 repeat pass, September 9, 2026, completed as specified, with a recorded workaround; observations confirmed by the driver.

• Repeating the same assignment on a second job took about a minute start to finish, roughly half the time of the first attempt, largely because the navigation was already known.
• The assignment control did not always show the visit's current team member when opened — this time it appeared blank even though the job's own summary already listed one — but saving still correctly reflected the existing assignment.
• As in the first test, the account had a single user, so we assigned the account owner rather than a separate technician, to test the assignment workflow on its own terms.
• No email or notification was sent when the assignment was saved, and no option offered to notify a team member beyond an unticked checkbox.

Limitations observed

• Same single-seat trial limit as run 029 (STS-TEST-JOBBER-CORE-2026-029): the account had a single assigned seat and the fixture technician "STS-TEST Tech One" was not added as a second user

Capabilities exercised: dispatch.

STS-TEST-JOBBER-CORE-2026-013 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• In our standardized workflow test, adding a job note in this product meant exploring more than one part of the job record: a line item's own description field looked like a place for a note, but the real notes feature turned out to be a separate widget that only reveals a text box and a photo-attachment button once clicked into.
• Once we found the right control, typing a note and attaching a photo together took under a minute, and the saved result showed as an authored, timestamped entry that persisted correctly after navigating away and back.
• A brand-new job record we started but did not save was abandoned without any on-screen warning, and it did not appear later in the jobs list - unsaved work can be lost silently.
• No plan upgrade or paywall was required to add a note or a photo to a job.

Capabilities exercised: photo documentation.

STS-TEST-JOBBER-CORE-2026-014 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• In our standardized repeat pass, once we knew where the notes and photo control lived, adding a note and a photo to a job and saving it took about 50 seconds start to finish.
• The save required no extra step to confirm - the sidebar updated immediately to show the note as saved, authored, and timestamped.
• No plan upgrade, paywall, or interruption appeared during this repeat run.

Capabilities exercised: photo documentation.

STS-TEST-JOBBER-CORE-2026-031 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• We created an invoice for a scheduled job by choosing the client and then which of that client's jobs to bill; the product filled in the job's line item and price automatically, with nothing typed by hand.
• The save button offered a plain save separate from send options (email, text message, or collect payment), so the invoice could be saved without sending or emailing anything.
• The saved invoice showed the correct line item, quantity, and total, and the same total and an unsent (draft) status appeared again in the account's invoice list.

Capabilities exercised: invoicing.

STS-TEST-JOBBER-CORE-2026-032 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• In a repeat run for a second job, starting the invoice directly from the job's own menu skipped the client- and job-selection steps and opened an invoice already tied to that job.
• The invoice was reviewed, saved without sending, and confirmed correct in under a minute; the invoice list then showed both invoices created during this session with matching totals and unsent (draft) status.

Capabilities exercised: invoicing.

STS-TEST-JOBBER-CORE-2026-033 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• We found the payment-recording control inside the invoice's overflow menu rather than a dedicated button, reaching it about 40 seconds after starting to look.
• The payment method list offered ordinary real-world options (cash, check, card, e-transfer, PayPal) with no sandbox or test mode; a separate online-payments setup screen exists but was not needed to record the payment.
• Recording a manual cash payment took one form with the invoice and balance already filled in, and one click to save; the invoice's status updated to paid immediately with a confirmation message, with no further steps required.
• The paid status showed consistently everywhere we checked afterward: on the invoice itself, in the account's payment list, and in the invoice list's balance column.

Capabilities exercised: payments.

STS-TEST-JOBBER-CORE-2026-034 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• On the repeat pass we used the same steps as before with no new discovery needed, reaching the payment screen in about 16 seconds.
• The payment method list matched what we found the first time: ordinary options like cash, check and card, with no sandbox or test mode.
• Recording the second manual cash payment took the same one form and one click as the first, and the invoice updated to paid immediately with a confirmation message.
• Reopening the paid invoice showed a full breakdown of the payment against the total, confirming a zero balance on both the invoice and the account.

Capabilities exercised: payments.

STS-TEST-JOBBER-CORE-2026-035 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Setting up a recurring job uses the same job-creation screen as a one-off job, with a toggle to switch from a single visit to a recurring schedule.
• Matching an annual repeat pattern took several menus: a preset list, then a small dialog to set the exact interval, then a separate field to decide when the recurring series ends - Jobber does not offer "end after N visits" directly, only a count of days, weeks, months, or years, which we converted ourselves into the right number of years.
• Once saved, the recurring job listed each future visit individually going out ten years, so the resulting schedule could be checked at a glance, about three minutes after starting the job form.
• The job's invoicing frequency was tracked separately from its visit schedule, and Jobber had already created an upcoming invoice entry for the job before any visit had taken place.

Capabilities exercised: recurring jobs.

STS-TEST-JOBBER-CORE-2026-036 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Repeating the recurring-job workflow on a second record followed the same steps as the first attempt and completed faster, about two minutes from starting the job form to the saved job.
• The interval dialog opened already set to a yearly repeat this time, though the field controlling when the series ends still needed a unit correction before it matched the intended number of years.
• The saved job's full recurrence details - the yearly schedule, the ten future visit dates, and the separate monthly billing frequency - were all visible immediately after saving, with nothing to scroll past or dismiss.
• As on the first attempt, Jobber had already created an upcoming invoice entry for the job before any visit had taken place.

Capabilities exercised: recurring jobs.

STS-TEST-JOBBER-CORE-2026-015 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• The product ships several client-facing reminder and follow-up messages already turned on by default, rather than requiring one to be created from scratch.
• Finding the setting took about 40 seconds and passed through an unrelated company-profile screen before reaching the right page; the product flagged that this settings page is being replaced by a newer one, though the older page still worked correctly.
• Editing the reminder's timing took one dialog with two schedule fields and a message template, saved with a single click and no confirmation prompt.
• The reminder editor has no way to name the rule or aim it at one specific customer; it is an account-wide setting that applies automatically to every customer.
• A text-message version of the same reminder is offered only on a paid plan; the email version used here was available without upgrading.

Capabilities exercised: automation.

STS-TEST-JOBBER-CORE-2026-016 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Repeating the workflow, the product's account-wide default reminders again meant nothing needed to be created from scratch; a second, different default reminder was configured instead of reusing the one from the first pass.
• The repeat pass reached the settings screen faster than the first attempt, about 16 seconds, once its location was already known.
• The setting changed in the earlier pass was still in place when the page was reopened for this one.
• As before, editing the reminder's timing took one dialog and a single save click, with no confirmation prompt and no way to name the rule or target one specific customer.

Capabilities exercised: automation.

STS-TEST-JOBBER-CORE-2026-017 first-use pass, September 9, 2026, not completed as specified.

• We looked for a way to set up an automated review request tied to a completed job in this product's trial account; the path is under Marketing, in a section called Reviews.
• That section turned out to be a promotional overview of the feature - sample screenshots and a short video - rather than a working settings screen; the toggles and buttons shown are part of the pictures, not clickable controls.
• The account's own connections settings showed that turning the feature on requires connecting a Google Business Profile first, which was not connected on this account; we did not make that connection during testing.
• We also checked the account's separate automation settings for quotes, invoices, and requests, and found no equivalent review-request option listed there.
• Finding this explanation took a wide search of about four minutes, including a web search, before landing on the actual reason the feature could not be configured.

This run is not admissible as a benchmark result: the fixture was not completed as specified (the run table has the row). Its observations above stand as recorded; it contributes no timing and no completed scenario.

Limitations observed

• A functioning review-request configuration screen: Marketing → Reviews (`secure.getjobber.com/reviews`) renders only a promotional/demo page (mock-up screenshots and a testimonial video) with no real form, toggle, or "Connect" control of any kind
• A connected Google Business Profile: Settings → Account Connections lists it "Not connected"; Jobber's own help content says the review-request automation is configured only after that connection is made, which the sandbox rules do not allow us to make

STS-TEST-JOBBER-CORE-2026-018 repeat pass, September 9, 2026, not completed as specified.

• This repeat pass confirmed the same result as the first: the review-request feature still could not be configured because the account's Google Business Profile was not connected.
• No new screens were checked in this short pass; it re-confirmed the account state left by the previous attempt.

This run is not admissible as a benchmark result: the fixture was not completed as specified (the run table has the row). Its observations above stand as recorded; it contributes no timing and no completed scenario.

Limitations observed

• Same as run 017: no functional review-request configuration screen, and no connected Google Business Profile (Settings → Account Connections: "Not connected")

STS-TEST-JOBBER-CORE-2026-037 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• The customer's own profile page doubled as the service history: opening it showed the customer's jobs, quotes and invoices on one page, with no separate screen to find and no extra click needed.
• Locating the customer and viewing their full job and invoice history took under a minute from the start of the task.
• Past invoices showed a paid status and a zero balance directly next to the job list, without switching screens.
• No help documentation, search, or guidance beyond the product's own navigation was needed to complete this task.

Capabilities exercised: crm.

STS-TEST-JOBBER-CORE-2026-038 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• On a repeat visit, the same customer profile again served as the complete service history, combining jobs, quotes and invoices on one page.
• The history view offers filters that can narrow the list to a single record type or expand it to show every record together, which surfaced additional past work beyond what was initially visible.
• Confirming that a specific job and its invoice were both present took well under a minute once the profile was open.
• No help was needed and no plan restriction or paywall appeared for this task.

Capabilities exercised: crm.

STS-TEST-JOBBER-CORE-2026-019 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Producing a client-data report took three steps: opening the reports section, choosing the clients report, and clicking export.
• The export did not download instantly in the browser; instead the product displayed a message saying the report would arrive by email within about 15 minutes.
• No account upgrade or paid plan was required to reach or run this export on the trial account.
• The emailed export arrived after the session and contained the expected client records; the read of the file was done off camera.

STS-TEST-JOBBER-CORE-2026-020 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Repeating the same export about a minute later followed the identical three-step path: reports, clients report, then export.
• As before, the product did not provide an instant in-browser download; it displayed the same message that the report would be emailed within about 15 minutes.
• No account upgrade or paid plan was required.
• The emailed file that arrived matched the earlier export exactly.

STS-TEST-JOBBER-CORE-2026-021 first-use pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• We opened Jobber in a resized browser window using Chrome's mobile-emulation mode and, about 45 seconds in, walked its core screens: the client list, a customer's profile, the jobs list, and a job record.
• The mobile layout replaces the desktop's always-visible side menu with a single "Menu" button; every screen we opened, about a minute and a half to two and a half minutes into the run, restacked into one column without cut-off text or overlapping buttons.
• On the job record, about three minutes in, an existing text note and its attached photo displayed correctly in the narrow layout, and the control for adding a new note or photo was present and usable.
• We did not test Jobber's dedicated schedule/calendar screen or compare what the mobile view can do against the desktop view in this pass.

Limitations observed

• Not a product limitation: the walk itself did not cover the dedicated Schedule/calendar screen or a desktop-vs-mobile comparison (step 4), so those two facts are recorded as "not on camera" rather than as gaps in Jobber

Capabilities exercised: mobile or responsive web workflow.

STS-TEST-JOBBER-CORE-2026-022 repeat pass, September 9, 2026, completed as specified; observations confirmed by the driver.

• Repeating the same walk a few minutes later, the mobile layout behaved the same way: a single "Menu" button in place of the desktop side menu, with every screen we opened stacking into one column.
• On a job record's note area, about a minute and 15 seconds into the run, clicking the photo-attachment button opened the computer's file browser, confirming the control actually triggers a file upload rather than being decorative.
• The customer's profile, reached about 40 seconds in, and a job's status, total, and existing note, reached about a minute and 10 seconds in, all displayed clearly on the narrow layout.
• As in the first pass, we did not open the dedicated schedule screen or compare mobile against desktop capabilities.

Limitations observed

• Not a product limitation: the walk itself did not cover the dedicated Schedule/calendar screen or a desktop-vs-mobile comparison (step 4), so those two facts are recorded as "not on camera" rather than as gaps in Jobber

Capabilities exercised: mobile or responsive web workflow.

STS-TEST-JOBBER-CORE-2026-023 first-use pass, September 9, 2026, completed as specified.

• We opened the app's integration marketplace and reviewed the QuickBooks Online and Xero listings without connecting either one.
• QuickBooks Online's listing named no plan requirement beyond needing a separate QuickBooks account, while Xero's listing stated it is available only on Jobber's Connect or Grow plans.
• Zapier automation is offered through the product's own settings page, with pre-built workflow templates, and links out to Zapier's site for the full list of supported triggers and actions.
• Searching the marketplace for a public API returned unrelated apps rather than a developer API listing, and no API or developer section was found elsewhere in the account's settings.

Limitations observed

• A public/developer API: searching "api" in the App Marketplace returned three unrelated apps (Trackem GPS, Ada – AI Email Marketing Employee, MindCloud); no developer/API item exists under Settings in this account
• The Xero integration is gated to the Connect or Grow plans per its marketplace listing; it was inspected, not connected

Capabilities exercised: accounting integrations.

STS-TEST-JOBBER-CORE-2026-024 first-use pass, September 9, 2026, not completed as specified.

• This run ran out of time when the trial account expired partway through, before the product's online-booking settings could be located.
• We searched the account's settings menu for an online-booking configuration screen but did not find one clearly labeled as such in the time available.
• After the trial ended, the plan-selection page we were redirected to did mention online booking as an included feature, though we did not get to verify it from inside a working account.

This run is not admissible as a benchmark result: the fixture was not completed as specified (the run table has the row). Its observations above stand as recorded; it contributes no timing and no completed scenario.

Limitations observed

• Online Booking's own configuration/preview screen was never located before the trial expired; no Settings item is labeled "Online Booking" (closest candidates in the sidebar: "Requests and Bookings", "Automations" — neither was opened)
• Time: the trial account itself ran out mid-run, before steps 2–5 could be attempted at all