Every customer, every brand, under one DOME.
iDome Systems gives large enterprises Retention Intelligence built on a Business Data Platform. It brings events from every brand, market and channel together, recognises each customer across all their accounts, and shows who is about to leave while there is still time to keep them.
Retention Intelligence, built on a Business Data Platform
Retention models are only as good as the data under them. In most enterprises the same customer is split across brands, markets, apps and older systems. iDome first builds one trustworthy record of what happened, to whom and when. Retention Intelligence then reads that record to find the customers who are drifting away and to start the right response.
One record of every event, across the enterprise
- Every sale, visit, payment, ticket and message lands as an event, so history can always be replayed and audited.
- Each person is resolved across every account, brand and channel they use.
- Each brand, market or business unit can be kept in isolation, through the same APIs.
- Every product, dashboard and agent works through the platform's APIs, never straight against the database.
Know who is leaving before they leave
- Activity is measured against each customer's own normal pattern, over three time windows.
- A customer moving to a new account counts as a move, not as a loss and a new sign-up.
- Consent and compliance rules run first. A failed check blocks contact and escalates.
- Retention actions fire at whichever level the score falls below its threshold: a customer, a segment or the whole base.
From raw events to the right action, in six steps
Capture
Events arrive from your systems through connectors and APIs, and are stored as they happened.
Classify
Each event is typed and labelled so it can be compared across sources.
Resolve
Records are matched into one person and one account, with rules designed for each business.
Enrich
Local AI models add context, such as intent, topic and similarity to past behaviour.
Score
Every journey stage and every customer gets a score against its own target or norm.
Route
Exceptions go to the owner, and the agreed workflow starts.
Find the quiet ones while there is still time
Most churn is visible weeks before it happens: fewer active days, a longer gap, a move to another account. Retention Intelligence watches for it per person, and acts only when the evidence is there.
Activity, in three time windows
Weekly, monthly and lifetime activity, each compared with the customer's own norm and blended with the health of the things that drive it. What counts as quiet differs from person to person.
One person, many accounts
Identity resolution links a customer's accounts using rules designed for your business. An old account going quiet next to a new one is a migration, not a lost customer.
Consent and compliance come first
Do-not-contact, cooling-off and similar flags carry across every account of a person. A failed check stops the offer and escalates to the compliance owner instead.
Drill from the whole base to one person
Start at the consolidated score, open a segment, then a single customer, and back again. Change a threshold and see which scores move.
Bars show active days against this customer's own norm (the dark mark).
A personal re-engagement message goes out today. If there is no activity in 7 days, the account owner gets a call task.
Know everything. Act on what matters.
The platform rolls every score into a Balanced Scorecard, so leadership sees the whole business through four perspectives. Then management by exception narrows that view to the few items that need a decision today.
Are we growing profitably?
Revenue and margin per customer, segment and channel.
Are customers staying?
Retention, re-activation and value, from the same scores Retention Intelligence uses.
Are processes on target?
Every journey stage scored against its own target.
Are we improving?
Data coverage, automation and how fast exceptions are closed.
While a measure is within tolerance, the platform stays quiet. When it breaches, it raises an exception with an owner, a severity and a workflow already started, and tracks it until the measure is back on target.
See the whole platform at work, from raw data to the action it starts
Two worked examples. All names, figures, targets and thresholds are illustrative.
A multi-line insurance agency places personal and commercial policies with several insurers. Each household may hold motor, home and life policies, bought and serviced through different people and systems. The agency wants to keep more households at renewal, and to know which part of its own process is losing them.
Identity resolution links every policy, insurer reference and contact detail to one household. Each stage of the journey is a business process with its own target and score (0 to 100).
Quote
quote returned within 24 h · actual 30 h72 WatchBind and onboard
documents complete at bind 95% · actual 97%91 On targetPremium billing
instalments on time 96% · actual 89%64 WatchService and changes
mid-term changes within 2 days · actual 1.688 On targetClaims
FNOL to settlement 30 days · actual 4158 At riskRenewal
reviewed 60 days ahead 100% · actual 82%70 Watch
Every process score rolls up into four perspectives
The scorecard shows the cause behind the customer number: slow claims and missed renewal reviews.
- Open claim past target (41 days)−1.8
- Two late instalments this term−1.2
- Renewal premium up 14%−1.0
- Three policies held, 6-year tenure+1.6
The agency principal sees three items, not three hundred
Every other measure is within tolerance and stays quiet.
- Consent and do-not-contact checks pass for every person in the household.
- The claim is escalated, and the household gets a status call from the claims handler within 2 days.
- The account manager gets a review task with the household's full history and the score drivers.
- Alternative quotes are prepared 45 days before renewal, and a payment plan is offered.
- The score is tracked to renewal, and the outcome is measured against a control group.
An online retailer sells through its storefront and a drop-ship platform. Suppliers ship directly to customers, payment processors pay out net of fees, and the books are kept in accounting software. Reporting was a monthly spreadsheet that took days and still could not say which products, suppliers or customers made money.
Order placed
captured from the storefront within 1 min99 On targetSupplier order
PO sent to supplier within 1 h · actual 40 min94 On targetShipped
supplier ships within 2 days · actual 3.455 At riskDelivered
delivered within promise 95% · actual 91%73 WatchPayout settled
payout matched to orders 99% · actual 97.5%78 WatchBooked
posted to the ledger within 1 day · actual 0.496 On target
True margin after fees, shipping, refunds and ads
| Product line | Revenue | Supplier cost | Fees and refunds | Ads | Margin |
|---|---|---|---|---|---|
| Home storage | 48,200 | 24,100 | 3,900 | 6,800 | 13,400 |
| Kitchen tools | 31,500 | 16,900 | 2,700 | 5,200 | 6,700 |
| Outdoor lighting | 22,800 | 13,300 | 4,600 | 6,100 | −1,200 |
| Total | 102,500 | 54,300 | 11,200 | 18,100 | 18,900 |
Each row is built from order-line events matched to supplier bills, payouts and ledger entries, so it reads the same at 10 products or 10,000 and across any number of stores.
The month-end close uses the same reconciled data, so the books and the dashboards agree.
Your data stays in your environment, and every change is traceable
Runs in your environment
The platform is deployed on infrastructure you control, on premises or in your private cloud. Customer data does not need to leave it.
Works with your data estate
iDome reads from the warehouse, data lake, CRM (customer relationship management) and operational systems you already run. It adds identity and retention scoring without replacing them.
Local AI models
Classification and enrichment run on local models, so there is no per-call cloud AI cost and no data sent to an AI provider.
Separation by brand and market
Data for each brand, market or legal entity can be kept apart in storage and in every API, to match your data protection and residency rules.
Event-sourced history
Every state the business cares about can be rebuilt from its events, which makes audits and corrections straightforward.
Every change traced
Each change to code, schema or configuration is tied to an approved enhancement record, so any line can be traced back to the decision behind it.
Large enterprises with many brands, markets and customer accounts
iDome is built for enterprises that serve the same customer through several brands, markets, apps and stores, where retention is worth the most and the data is hardest to join up.
Start with one brand or market, then roll out
Briefing
Your retention goals, brands, markets and the systems that hold customer data.
Assessment
We map your data estate and agree the identity rules and what an active customer means for you.
Pilot
Retention Intelligence on one brand or market, on your own data, in your environment.
Prove
Scores and actions are compared against a control group, so the effect is measured.
Roll out
More brands, markets and channels on the same platform and identity model.
See your own customers under the DOME
In a briefing we show Retention Intelligence and the platform on a sample shaped like your business: your brands and markets, your systems, your identifiers and your idea of an active customer.