Written by the architects and consultants doing the delivery. Opinionated where we have earned it, specific where it matters, and free of the release recap you can already read on the vendor site.
Latest · 59 articles
Field note11 min
The automation argument is never really about Flow or Apex. It is about who maintains the thing in eighteen months, and whether anyone can reconstruct why it exists.
Playbook11 min
Twelve pre-flight checks we run before an AI agent is allowed near a real customer, each with the failure it catches and the fix. A go or no-go list, not a readiness score.
Decision guide12 min
Eight decision-stage questions to settle before a Salesforce CPQ migration is scoped, each with what a good answer sounds like and what a bad one sounds like.
Architecture11 min
Why the tool gets chosen first, what that order costs once the project has closed, how to run the decision the right way round, and how to have the conversation when the platform is already bought.
Org health11 min
Access sprawl is the most common form of org debt and the least visible until an audit. How it accumulates, how to measure it, and a repeatable way to clean it up without breaking Monday.
Architecture11 min
Most of an org is reversible. A handful of early modelling choices are not, and they get made in the first fortnight by whoever is fastest to a whiteboard.
Industry playbook12 min
A regulated organisation changes the data model, the access design, the pace of release and the definition of done. What a bank, an insurer, a hospital and a utility all do to a Salesforce delivery, and what to plan for before the first workshop.
How we work11 min
A repeatable thirty-minute method for turning Salesforce release notes into a short, org-specific action list: the preparation that makes it possible, the four buckets every item lands in, and what to test in the preview window.
Delivery20 min
Evaluating a Salesforce partner on evidence rather than credentials: what AppExchange really tells you, the questions that separate delivery from a pitch, and the contract terms that matter.
Operations12 min
A decision framework for Salesforce managed services against hiring in-house: the variables that actually decide it, the real failure modes on both sides, and the hybrid most organisations end up needing.
Security16 min
Salesforce decides record visibility in a fixed order. Understanding that order is the difference between fixing access and layering more of it on top, and it is where audits go wrong.
Engineering12 min
Page load, list views, reports and large data volumes: where Salesforce slowness actually comes from, how to measure it before changing anything, and what to fix first.
Operations11 min
An ordered technical debt audit for a Salesforce org: what to inspect in each area, what a bad result actually looks like, and the question that decides whether the debt is worth paying down.
Delivery13 min
Ten Salesforce implementation mistakes ranked by reversal cost rather than by how bad they look, with the cheaper decision that was available earlier in each case.
Architecture11 min
Each Salesforce integration pattern characterised by what it assumes about the far side, how it behaves under volume, the consistency it gives, and the condition that rules it out.
Engineering14 min
The code-level and contract-level decisions behind a durable Salesforce integration: API family, authentication failure modes, limits as design constraint, idempotency, retries, partial failure and versioning.
Architecture11 min
System, process and experience APIs mapped onto real Salesforce objects: how to scope a System API, where business logic is allowed to sit, and when the third layer is unnecessary.
Architecture10 min
The integration decisions taken above the project line: attribute-level ownership, who funds an interface for its life, the register of what already connects, and portfolio sequencing.
Field note11 min
The Flow mistakes that survive testing: correctness faults organised by the condition that exposes each one, with the production symptom, the cheap confirmation test, and the fix.
Analysis10 min
The Flow or Apex criteria applied to ten real requirements: where each one crosses into code, where volume moves the line, and where the honest answer is neither tool.
Guide11 min
A measurement-first guide to Salesforce Flow performance: how to tell a slow flow from a slow transaction, the faults that actually cost time, and the evidence for each.
Architecture11 min
Naming, ownership, one record-triggered flow per object, entry criteria, ordering and a deprecation path: the Flow conventions that still work once the estate outgrows the person who built it.
Field note11 min
The symptoms Revenue Cloud programmes present when they are in trouble, the lifecycle stage each cause actually lives in, and the first move that confirms or eliminates it.
Playbook11 min
A delivery playbook for a Salesforce CPQ to Revenue Cloud migration: scope splitting, the catalogue rebuild, price rule intent, parallel running, cutover timing and in-flight quotes.
Analysis11 min
CPQ covers configure, price and quote. Revenue lifecycle thinking runs through contract, order, billing and recognition too. A stage-by-stage comparison, and how to locate where your own complexity sits.
Architecture11 min
The Revenue Cloud decisions that constrain everything after them: what a product is, which prices are calculated rather than negotiated, bundle depth, and how amendments behave.
Analysis8 min
Build or configure, argued over five years rather than one. The line items teams omit, the three cases where a custom portal genuinely wins, and what you own the day you build your own.
Playbook11 min
Task completion is the only portal metric worth optimising. How to find the real top tasks, why traffic misleads, where login loses people, and the honest deflection test.
Architecture12 min
External users are a different sharing problem from internal users. Licence types, sharing sets, the guest user, and the external access design checklist to settle before anyone picks a theme.
Playbook12 min
Worked service automation examples by case lifecycle stage: what each replaces in the agent workflow, the trade accepted, and the automations we removed again.
Playbook10 min
How to design Omni-Channel from measured capacity rather than assumed capacity: work item weighting, queues that model skill, when skills-based routing is over-engineering, and why the overflow path deserves designing first.
Guide11 min
Why a taxonomy built from what the business wants to report fails, why the catch-all value destroys the dataset, and how to migrate a live taxonomy without losing historical comparability.
Guide11 min
Design principles for Service Cloud automation: what should decide, what should only notify, why notification rots, and how to retire rules faster than they accrete.
Playbook11 min
Worked pipeline automation examples by stage: what each one does, what it costs in flexibility, and the automations we have taken back out again.
Guide12 min
Opportunity stages defined on buyer evidence: exit criteria per stage, what to do with probability percentages, close-date slippage, and changing a stage model safely.
Guide11 min
Lifecycle stages both departments would define the same way, the lead-to-account boundary, routing that holds under volume, and a recycling path for the leads you rejected.
Field note12 min
A diagnostic for low Sales Cloud adoption: the six causes behind it, the signal each one leaves in the org, and the intervention that actually fits.
Guide11 min
Design principles for Sales Cloud automation: who each rule actually serves, the anti-patterns that produce workarounds, and where automation should stop.
Analysis11 min
Where generated marketing content needs a human gate, and why the answer follows from how reversible the channel is rather than from how good the model has become.
Guide11 min
Why marketing estates become untouchable rather than broken, and the governance that prevents it: naming, ownership, parameterisation, centralised suppression and a retirement policy.
Playbook15 min
A sequencing playbook for Marketing Cloud Next: the six early decisions that determine whether the estate is still workable at eighteen months, ordered by what is expensive to revisit.
Decision guide8 min
Growth Edition is published per org rather than per user, which takes headcount out of the argument. What is left is a capability threshold, and a short list of questions about your own operation that finds it.
Analysis10 min
Marketing Cloud Next and Marketing Cloud Engagement are separate platforms that coexist under the Agentforce Marketing brand. What each is, who each suits, and why a move between them is a migration.
Field note11 min
The six places Data 360 programmes stall once they are already underway, the observable signal for each one, and the recovery order that works: freeze scope, restore one narrow profile, then widen.
Architecture9 min
The path from source system to agent answer, and the four points along it where grounding quality is actually decided. Written for teams who have tried to fix a wrong answer with a better prompt.
Analysis7 min
A warehouse is built for analysis and Data 360 is built for activation. Where each fits, what breaks when you swap them, and why Zero Copy made keeping both cheaper than it used to be.
Analysis11 min
A prerequisite list rather than a wish list: realistic Data 360 use cases mapped to the source systems and the specific data quality each one demands before it works.
Architecture11 min
The five Data 360 decisions that are cheap before the first connector runs and expensive afterwards: scope of unification, identity resolution match rules, source of truth per attribute, copy versus reference, and ownership.
Field note11 min
A symptom-first diagnostic for Agentforce deployments: ten distinct implementation failures, the observable behaviour each one produces, and how to confirm which one you actually have.
Architecture15 min
Your sharing model is the AI security boundary. How the running user decides everything, why composition changes exposure without changing permissions, and a pre-launch checklist for customer-facing agents.
Analysis10 min
A fourth test for sales AI candidates beyond value, data readiness and error cost: workflow fit. What genuinely helps reps, what they route around, and why review is not a control.
Playbook9 min
How to measure service automation honestly, why deflection alone is a vanity metric, and an escalation design framework covering triggers, routing and carried context.
Architecture9 min
A reference architecture for Agentforce on Service Cloud: how channels reach the agent, what it can read at each hop, where the case object sits, and how escalation carries context to a human.
Decision guide9 min
The three published Agentforce buying models explained, plus a cost model covering the grounding, permission and evaluation work that usually exceeds consumption spend in year one.
Analysis7 min
A decision table separating work that belongs in deterministic automation from work that justifies an agent, and why most of what happens in a Salesforce org should stay in Flow.
Analysis9 min
A scoring matrix for Agentforce use cases across value, data readiness and failure cost, including the popular ones that consistently fail to justify themselves.
Playbook12 min
A phase-by-phase Agentforce implementation runbook with explicit entry and exit criteria for each gate, written for the team who has to deliver it rather than the one buying it.
Architecture15 min
The five readiness dimensions that decide whether an Agentforce deployment survives contact with real customers: scope, grounding, permissions, escalation and evaluation. A framework you can score your own rollout against before you commit a budget.
The same thinking, applied. Each one names the products involved, shows the architecture, and says what was contested along the way.
No newsletter yet, we would rather publish well than publish on a schedule. Subscribe to the feed to get pieces as they go up, or bring us a specific problem and we will answer it directly.
