Engineering
A CRM nobody updates
is an expensive address book.
Adoption is the only CRM metric that predicts every other one. We design around what a salesperson will actually do between two meetings on a phone, then automate the data entry that they were never going to complete anyway.
Overview
Salesforce, Dynamics or a custom build — chosen on your data model, not our licences.
Start with the record that matters. In a lending business it is the application; in industrial sales it is the quotation; in a hospital group it is the referral. Modelling the CRM around the object your business genuinely transacts, rather than around the vendor's default opportunity pipeline, decides whether the system gets used or worked around.
Platform choice comes second and is genuinely open. Salesforce and Dynamics 365 bring a large amount of finished behaviour and an ongoing per-seat cost that scales with headcount. A custom build on your own stack costs more upfront and less per year, and it makes sense when your process is unusual or when seat counts run into the thousands.
Whatever the platform, the same three things determine adoption: the mobile experience for people who are not at a desk, the quality of the data that arrives without anyone typing it, and how quickly a manager can answer a question without exporting to a spreadsheet. We build against those three before we build reports.
- 91%
- weekly active use at Tinsmith Retail, 90 days after rollout
- 23 min/day
- manual data entry removed per rep through activity auto-capture
- £186k
- annual per-seat licence avoided by moving 640 users to a custom build
Capabilities
What this covers
Six areas we staff properly. If your problem sits outside them, the honest note at the foot of this page says so.
Data model and process design
Objects, lifecycle states and permission boundaries modelled on how your business actually transacts. Stage definitions get exit criteria that a system can check, so pipeline reports mean something.
Salesforce and Dynamics 365 implementation
Configuration first, Apex or plugins only where declarative tooling genuinely cannot reach. We keep an inventory of every code extension with its trigger, owner and upgrade exposure.
Custom CRM builds
When per-seat licensing or an unusual process makes packaged CRM the wrong shape, we build on your own stack with the field-level security, audit trail and offline behaviour you would expect from a vendor.
Telephony, email and messaging integration
Click-to-dial, call recording, mailbox sync and WhatsApp Business threading wired into the timeline, so activity capture happens as a side effect of the work rather than as homework.
Field and mobile enablement
Offline-capable mobile apps for reps and engineers, with conflict resolution on sync and forms short enough to complete in a car park between two site visits.
Migration from spreadsheets and incumbent CRMs
Deduplication with survivorship rules the sales leadership signs off, history preserved where it is worth preserving, and a documented decision on what deliberately does not come across.
Deliverables
What you get
- Entity and lifecycle model with stage exit criteria defined
- Configured or custom CRM with a role-based permission matrix
- Telephony, mailbox and messaging integrations with activity auto-capture
- Offline-capable mobile app for field and sales teams
- Migrated and deduplicated records with survivorship rules documented
- Manager dashboards and a training pack aimed at daily users
Stack
What we build it with
- Salesforce Sales Cloud
- Apex
- Microsoft Dynamics 365 Sales
- Power Automate
- HubSpot
- Next.js
- NestJS
- PostgreSQL
- Twilio
- React Native
- Elasticsearch
- Metabase
Process
How the engagement runs
Two-week increments against a written definition of done. You can stop at any increment boundary and keep everything built so far.
Shadow the users
A week riding along with reps, service agents or field engineers, watching what they record, what they avoid recording, and what they keep in a private spreadsheet.
Model and choose
The data model is designed first, then platform selection is scored against it on licence cost at your headcount, extension limits and integration fit.
Build the first territory
One team, region or product line goes live within eight weeks, so adoption assumptions get tested on real quota-carrying users before the rollout scales.
Roll out and migrate
Remaining teams onboard in waves with data migrated per wave, each preceded by a dry run and followed by a two-week usage review against adoption measures.
Tune on usage data
Field-level telemetry shows which inputs nobody fills in. Those fields get removed, defaulted or automated rather than mandated through a policy email.
When this is the wrong engagement
If your sales operation is under about fifteen people and closing deals from a shared inbox, a CRM programme will slow you down before it helps.
FAQ
Questions we get asked
- Salesforce or a custom build?
Run the arithmetic at your five-year headcount. Packaged CRM wins on time-to-value and on the breadth of behaviour you get for free; a custom build wins when seat counts are large, the process is genuinely unusual, or data residency rules make the platform awkward. We deliver both.
- Our last CRM rollout failed. What would be different?
Usually the fix is subtraction. Failed rollouts almost always ask reps to type things that produce no value for them, so we cut mandatory fields hard, auto-capture activity from telephony and email, and use weekly active use as the release criterion rather than configuration completeness.
- Can you integrate CRM with our ERP?
Yes, and the direction of truth needs deciding before any code is written. Customer master, pricing and credit status normally originate in ERP; pipeline, activity and forecast originate in CRM. We write that ownership down per field, then build the sync with reconciliation on both sides.
- How long until it is live?
Eight to twelve weeks for a first territory or team, including migration of their records. Full rollout depends on how many distinct processes exist across your business units; the mistake is trying to unify all of them before anyone has used the system in anger.
Related
More in Engineering
Engineering
Blockchain Solutions
Permissioned ledgers, tokenised assets and independently audited smart contracts.
Engineering
IoT Development
Firmware, connectivity and telemetry pipelines for devices you cannot easily reach.
Engineering
Legacy Modernisation
COBOL, Delphi, WebForms and Oracle Forms moved off, without a big-bang cutover.