Larkspur Works mobilisation report

Larkspur Works
mobilisation report

The people are keeping the work moving. The organisation is making that harder than it needs to be.

Fictional organisation. Illustrative data. No client information. Larkspur is an invented repair, training and facilities enterprise: 124 people, three sites and 31 tools. This is an illustrative example, not a client case study.

Larkspur needs the capability to carry change through

It lacks a joined-up way to define, prioritise and deliver strategically useful technology change. No end-to-end request owner, 13 unowned tools and uneven protected time point to an organisational gap. This is a judgement about how change is led and supported, not a verdict on individual technical talent. [E4–E5, E8–E10]

An email becomes a planning-sheet row. Missing details trigger a call; a helpful colleague gets the job moving. Finance later reconstructs a separate record. This invented scene illustrates the handover pattern. The spreadsheet itself is not the diagnosis. [E4, E8]

Our recommendation

Do not start with an estate-wide replacement. Give request-to-schedule one owner, agree the job pack and exception route, then test a shared queue using the existing tools. Resolve six approaching renewal decisions alongside that work, before notice dates lock in the options. [E4–E6]

The decision now: appoint the service owner and finance sponsor; consider the separately agreed delivery stage on section 16. Mobilisation diagnoses the problem and recommends the plan; it does not include that delivery automatically.

Powered by yopla.app. The £37,440 modelled resource cost and cash limitations remain on section 15. All figures and reconstructed scenes are invented; no client quotation or screenshot is reproduced.

What the evidence can support

The example combines broad participation with a smaller set of deeper checks. Coverage is visible so that absence of evidence is not mistaken for absence of a problem.

Participation is uneven

Customer support and finance are well represented. Leadership and support have the lowest response rate. Non-respondents may have different experiences; the 98 responses are not treated as a random sample or reweighted to imply one. Site and shift-level checks remain necessary before organisation-wide conclusions. [E1–E2]

Three different kinds of statement

Recorded means a count or observation in this fictional evidence pack. Modelled means an estimate calculated from stated assumptions. Proposed means a target, decision or future action. None represents a real measurement of a real organisation.

Not covered: penetration testing, accessibility certification, payroll validation, a representative time-and-motion study, supplier due diligence, or a psychological assessment. The document is a mobilisation example, not independent assurance.

The example also includes leadership and anonymous readiness instruments, maturity interpretation, cover and document checks, adoption reconciliation and enquiry outcomes. Their separate denominators and limitations are listed on section 18.

A shared direction, uneven foundations

Digital maturity is how well an organisation can choose, use and improve technology to support its purpose. It depends on the people, decisions and everyday work around the tools, as well as the tools themselves.

How Yopla makes the picture clearer

Our maturity framework brings the evidence together across six dimensions: Strategy & Leadership; People & Change; Operations & Delivery; Technology, Data & AI; Growth, Marketing & Revenue; and Sustainability & Environment. This keeps the diagnosis wider than a software problem and shows which foundations need to improve together.

We compare self-assessment, anonymous responses and Yopla’s evidence-based interpretation. Differences prompt questions, not verdicts on people. For Larkspur, the priorities connect: agree direction and decision rights, give the handover an owner, and protect time to learn before introducing more automation. Sustainability remains unassessed because the evidence is missing.

Self assessed: six sponsor/leadership respondents. Anonymous: 72 respondents, with 72 valid ratings on each of the five assessed dimensions. Yopla: explicit auditor interpretation of the fictional source pack. These are different perspectives and cohorts, not paired measurements. [E9–E10]

Scale: 0 Initial · 1 Emerging · 2 Defined · 3 Managed · 4 Optimising. Attained bands use the floor of the score: 1.8 is Emerging, not Defined. No whole-organisation average is calculated. Missing evidence is not zero.

The client-facing report has three tracks. Yopla uses an auditor score where supplied, otherwise an observed score; it never blends them. This example supplies auditor judgements for five dimensions, leaving sustainability unassessed. [E10]

Buying a tool will not assign an owner.

The evidence points to a system of work that depends on people repairing its gaps. Fix the connections, not just the screen.

Repeated entry, clarification, help nominations and cover gaps are recorded in the fictional sources. The chain above is an explanation to test through observation and an exception rehearsal, not a proven causal result. The cohorts overlap and must not be added together. [E2, E4, E7–E11]

New software changes habits, responsibilities and culture

A shared queue only works if people trust it, managers use it and the next team accepts its records. Agree the process, practise exceptions, give staff time and make concerns safe to raise. Otherwise the old spreadsheet and informal workarounds may continue beside the new tool.

What would change our mind?

Test the explanation. If clearer ownership does not reduce returns, investigate workload and scheduling. If a deputy cannot resolve an exception, revise the guide or decision rights. Build only where existing tools cannot safely carry the agreed work.

Owners: service owner tests the handover; sponsor resolves decision rights and time. Commercial and environmental evidence tasks remain separate.

Change follows the real network, not only the chart.

The shadow hierarchy is the route people actually use for help and decisions. Understand it before changing how work moves.

Recorded in the fictional survey: 31 of 74 cross-team help nominations (42%) point to three bridge roles: service coordinator, planning lead and finance analyst. Nominations are mentions, not unique people helped, hours of work or proof of overload. The diagram summarises role connections; line weight does not represent a measured workload. [E2]

The informal network will shape adoption

Three bridge roles attract 31 of 74 help nominations. People may be relying on them to cross unclear boundaries. Involve these trusted contacts in design, test cover and capture their exception knowledge. A manager announcing a new process will not automatically replace the route people trust. The nominations alone do not prove overload or resistance. [E2, E11]

Support the people who make the system work

Involve the bridge roles in the pilot with protected time. Pair each with a deputy and capture decisions in the shared workflow. Do not simply route more work to the most nominated person. Invite staff to correct the map before it is used to change responsibilities.

Privacy and interpretation: a public example uses role-level, fictional data. A real report should use an agreed access model and avoid identifiable small-group reporting. Network position is not a performance rating.

A helpful colleague is not a resilience plan

Look at who people turn to, what knowledge is accessible and whether another person can carry the work when the usual contact is absent.

The help network on section 5 is a nomination pattern. The cover check below uses a different source: a structured review of the 12 key workflows. A nomination does not establish dependency, workload or essentiality. [E2, E7, E11]

Follow knowledge, not just reporting lines

In this fictional review, three cross-team workflows require decisions from more than one function. Two have no shared exception log. Capture the reason for each exception and its resolution so the next person does not have to rediscover the route. Operations owns this work. [E11]

Make the safe route easy to find

Of the 98 named survey respondents, 62 could identify the agreed security contact; 36 could not in that instrument. This is a recognition check, not proof that an incident would be reported safely. The technology lead should publish one route, rehearse a scenario and confirm out-of-hours cover. [E2 extension]

All roles, connections and counts are invented. Public examples stay at role or workflow level. A real engagement requires agreed access and protection against identifying small groups.

People need room to change, not another instruction.

Most respondents understand the reason for change. Far fewer have time to learn it. A new system could feel like extra work unless managers change those conditions.

Favourable means agree or strongly agree on each item. Each item has 72 valid responses from 72 anonymous respondents; 52 of the 124 people in scope did not respond. Responses may overlap between items. No team-level breakdown or individual linkage is used. [E9]

Readiness is weaker than understanding

Understanding is favourable for 54/72 respondents; learning time for only 28/72, and safe challenge for 39/72. This suggests a risk of hesitant or surface-level adoption, not proof that the whole culture is change-averse. Managers must protect time, hear concerns and remove competing work. Only nine of 12 champions have time agreed. [E8–E9]

These three single-choice responses are an illustrative working-style summary, not a validated personality typology. They are separate from the learning-format question on section 8 (39 guided practice, 31 written guide, 28 peer demonstration). No cross-tabulation is implied. [E2 extension]

Preferences change how we introduce the work

Personality and operating-style profiles help anticipate who wants a clear plan, a practical trial or time to discuss uncertainty. Learn this before choosing the rollout and support. Offer those routes; check with each person what helps. A preference is not competence or a fixed identity. Anonymous culture answers must never be linked back to individuals.

Owner and check: People lead agrees protected time before the pilot; sponsor reviews concerns in aggregate. Repeat the same anonymous items after the first phase, reporting coverage and any change in who responded.

Teach the changed task, not just the new software.

Only 48 of 86 participants passed a complete workflow record unaided. A new interface will not remove that support need. Design the process, practice and instructions together.

Recorded illustrative results: the same 86 participants attempted five independent scenarios. Each bar counts people completing that task without assistance; a participant may appear in several bars. The scenarios are not a validated ability test and should not be used to rank staff. [E3]

Prioritise the tasks the change will require

The pilot depends on finding the correct shared record, checking data quality and handing work to the next owner. Start with those tasks. The lower safe-AI result suggests a separate support need; it does not justify making AI a dependency of the service pilot.

Preferences shape delivery, not judgements

Of 98 survey respondents, 39 selected guided practice, 31 a short written guide and 28 peer demonstration as their preferred first format. Offer all three, with accessible alternatives and time to practise. Preferences can change with the task and circumstances. [E2]

Operating style is not competence. Personality or preference data must not be used to infer ability, motivation, suitability for a role or performance. Confirm an individual’s needs with them. In this example, 12 people volunteer as champions; only nine have protected time agreed, so the plan assumes nine until the remaining capacity is confirmed. [E8]

Design for the practical effort the work requires.

Confidence can hide a support need; hesitation can hide capability. Check real tasks before deciding how much input, training and support the change will demand.

One fictional shared-record handover task, with a pre-task confidence question. High confidence means 4 or 5 on a five-point scale. The 80 matched participants are a subset of the 86 practical participants; six additional people completed the task but lack a matched confidence answer. Thus 42 + 6 = the 48 of 86 completions on section 8. [E3 extension]

Typing matters when the job depends on repeated entry

Typing helps reveal the effort of a keyboard-heavy task. Use it to question repeated entry, prefill known data and plan support. Device, accessibility and task context affect results; 12 people did not take the exercise. It is not intelligence, a universal speed target or a basis for individual productivity claims. [E3 extension]

Design support around the work

Offer guided practice to the 38 matched participants needing help, including 21 who felt confident. Give the 15 lower-confidence completers unfamiliar cases without labelling them incapable. People lead checks accurate handover and safe exceptions, not attendance or speed alone.

Fix the handover before automating it

The request-to-schedule journey is the first pilot because it combines visible service friction, repeated entry and a manageable boundary.

Clarification: incomplete information returns to the sender. Ownership gap: no visible next action when an exception occurs.

These counts come from the same 42-request fictional sample and may overlap. They must not be added to claim 37 separate failures. The timing subset contains 24 complete requests: median elapsed time is 3.8 working days; median hands-on administration is 38 minutes. These are different measures and cannot be subtracted to infer idle time. [E4]

What the sample suggests

Information is collected before the next team’s acceptance criteria are clear. Staff then repair the job record through email or personal contact. Adding automation now could move incomplete work faster without resolving why it comes back. The first design task is a common definition of a request that is ready to schedule.

Pilot design

Use one site and one request type for four weeks. Include routine work and a defined set of exceptions. Retain a fallback queue, reconcile it daily, and record every bypass. Compare like-for-like requests before deciding whether the change should spread.

Proposed outcome: fewer returned handovers and shorter elapsed time, with no deterioration in safety checks, customer communication or staff effort. The target and release criteria are on section 16.

One pilot, a wider improvement backlog

The request handover is the worked example. Mobilisation also exposes the quieter processes and missing evidence that shape delivery risk.

Can somebody else use the documents?

Instructions: six current, three outdated, two confirmed absent and one not received. Those categories sum to 12. Owner and cover counts overlap with instructions; they are not additional workflows. “Not received” is an evidence gap, not proof a document does not exist. [E7, E11]

A separate request list contains 16 policies or control documents: nine current, four overdue for review and three not received. This is a document-status check, not assurance that the controls work. Technology and People leads agree an owner and review date for each gap. [E12]

Prioritise by consequence and readiness

Keep the first pilot narrow. Work on access evidence and time-sensitive renewals in parallel. Put the remaining opportunities into an owned ledger with evidence, dependencies and a decision date. Do not add their modelled benefits together until overlap has been checked.

Make the estate governable

An inventory is only useful when it shows what each tool supports, who owns the decision and what happens when it changes or fails.

Ownership: 18 tools have a named business owner; 13 remain to resolve.

There are 24 paid agreements and seven tools with no direct subscription charge. Spend is the annualised committed amount in the fictional contract register, excluding VAT and internal support time. It is not total cost of ownership. The category bars count tools, not contracts or users. [E5–E6]

Three findings to resolve

Keep the useful capability

Map the shared queue, records, reporting and finance checks to what is already licensed. Test configuration and integration options before adding another product. Retention is conditional on support, usable access controls, exportability and an affordable way to maintain the connection.

Evidence limits: user-reported adoption is not login telemetry. A tool reported by several teams may still have low use or duplicated records. This example does not claim that every asset, integration, licence or security control has been independently verified. [E2, E5, E7]

Reported use is a starting point

Connect the contract register to how work happens. A licence, a survey mention and an observed task are three different kinds of evidence.

Reconcile before consolidating

The technology lead should resolve the four extra names with the reporting teams, identify the work they support and check information boundaries. Finance then confirms whether any paid commitment exists. No shadow-spend amount is added to the £286,400 register without evidence. [E2, E5–E6, E13]

Accountability: Technology lead owns reconciliation; finance owns commitment evidence; service owners accept any change to their work. Popularity does not establish criticality, and a missing mention does not establish non-use.

All adoption and tool-name counts are invented. The six observed tools are a selected subset of the eight checked-in-scope tools and the 31-tool register. There is no usage telemetry or security certification in this example. [E13]

Use the renewal window to make decisions

Six fictional agreements represent £74,200 of annualised spend and renew within 90 days. Decide against service needs and notice dates, rather than treating renewal as an automatic purchase.

Day 0 is the decision baseline at the end of this illustrative mobilisation. Delivery requires a separate agreement; supplier deadlines still apply. Dates and terms are invented. Spend in this window is 25.9% of the £286,400 register; the other 18 agreements total £212,200. Notice dates precede renewal dates and must be confirmed against the signed terms in real work. [E6]

No exit without a service test

For any consolidation, confirm which work and records rely on the supplier, validate a usable export, test the replacement with the people doing the work, and cost migration and ongoing support. Do not count an avoided renewal while a necessary overlapping contract remains.

Decision posture: preserve continuity now, challenge duplication with evidence, and avoid a long commitment that pre-empts the pilot. Supplier names are invented and are not recommendations.

Separate potential from money in the bank

The model makes a modest case for testing the change. It does not yet establish a positive first-year cash return or a reason to reduce staffing.

The two staff cohorts are disjoint. The 50 hours per week is a scenario assumption informed by fictional interviews, not a measured total from the 42-request sample. Licence candidates exclude the labour estimate. All ranges are scenarios, not confidence intervals or guaranteed savings. [E6, E8, M1]

A bounded investment case

The external allowance is a fictional planning assumption, not a Yopla quotation. Internal time is valued at 480 × £28 = £13,440. The resulting £37,440 is a resource-cost estimate; it is not all additional cash spending. Recurring support, price changes and any implementation tax treatment require separate confirmation.

Proposed first release: cap discovery and pilot preparation at £6,000 external spend plus 120 internal hours, within the total allowances above. Seek explicit approval for later phases after the first gate. These are example governance thresholds, not an instruction or commitment.

Do the hard work before the build.

Mobilisation supplies the diagnosis and recommended priorities. A separately agreed delivery stage should earn each release through evidence, not activity.

A small scorecard that people can challenge

Weekly review: inspect exceptions, workload, accessibility feedback and data quality alongside the scorecard. Escalate a missed renewal deadline immediately. Do not expand if safety, service continuity or information handling deteriorates, even when headline speed improves.

Day 90 decision: continue, revise or stop based on observed outcomes and total cost. There is no assumed organisation-wide software rollout in this plan.

Trace each finding to evidence

A real mobilisation report should let its reader trace each material finding to evidence, understand the limits and see what decision comes next.

Definitions that prevent misleading conclusions

Coverage is the share of the defined scope represented in a source, not proof that its answers are correct. Current instructions means reviewed against the task and usable by a second person. Annualised spend converts recorded commitments to a yearly rate; payment dates can differ. Recovered capacity is time made usable for other work, not a cash saving.

What would be needed in a live engagement

Replace every synthetic source with permissioned evidence, verify its date and scope, reconcile contradictions with staff, confirm supplier terms, and agree who may see person-level data. Have owners review the findings and record unresolved challenges. Keep a versioned decision log in yopla.app as work proceeds.

The decision requested in this example

Consider a separately agreed, bounded first delivery phase; appoint the service owner and finance sponsor and agree the Day 14 gate. Decide later phases from pilot evidence. This demonstrates the output of mobilisation: a clear starting point, transparent reasoning and a practical way to deliver and verify change.

Illustrative mobilisation example. Powered by yopla.app. All organisation and supplier data is fictional. No client findings, quotations or screenshots are reproduced.

The source behind each conclusion

The diagnosis is built from a defined set of fictional records. Keep the source, its coverage and its limits attached to the finding.

Extensions to existing synthetic sources

E2: the same 98 named respondents supply working-preference choices (37/34/27) and security-contact recognition (62/36). These are distinct questions. E3: 80 matched confidence answers within the 86 practical participants; 74 optional typing exercises. E7: the same 12 workflows now show current/stale/absent/not-received status. Existing source counts and economic assumptions remain unchanged. E4/E8: the opening email, planning-sheet and finance-record scene is an invented reconstruction of the handover pattern, not a quotation or an additional measured case.

About this example

Larkspur Works and every person, supplier, figure and connection in this report are invented. The separate Northfield programme illustrates a different fictional organisation. These examples show the approach; they do not describe a client or promise the same results.

Mobilisation output: a diagnosis, audit trail, priorities and recommended next-stage plan. Pilot delivery, integration, training and operational handover require a separately agreed next stage. Powered by yopla.app.