Getting started with Warranlytics: your first week

Most warranty systems fail in the first fortnight, and almost never for technical reasons. They fail because someone tried to import four years of history before anyone had registered a single live sale, ran out of patience halfway through, and ended up running two systems at once.

So this is a first week designed around a different order: get today's sales captured correctly, prove one claim works end to end, and leave the archive until you know the process holds.

Before you start: the pre-flight checklist

Half an hour with these answered will save you a week of rework.

• Your product list. The twenty or thirty lines that account for most of your warranty exposure. Not the full catalogue.
• The warranty terms for each. Coverage period, what is included, what is excluded. If these only exist as a paragraph on an invoice template, write them down now.
• Your serial number situation. Do your products carry serials? Are they unique? Are they printed somewhere a person at a counter can actually read? If the answer to any of these is no, sort out serial number tracking first — it is the key everything else hangs on.
• Who registers warranties. Named people, not "the sales team".
• Who approves claims. And who can override a rejection.
• What you will do with the archive. A decision for later, but know now whether the historical data exists in a usable form at all.
• Your current claim volume. Roughly, per month. It determines which plan you need and when.

If you cannot answer the terms question, stop and read how to write warranty terms before you configure anything. Terms you cannot apply consistently will make every claim a judgement call forever, whatever software you run.

Day one: create the organisation and see the shape of it

Signing up creates your organisation — the container everything else sits in. Products, warranties, customers, claims, and users all belong to it, and access is role-based within it.

Spend the first session looking rather than building. Open each area, work out what a warranty record holds, what a claim record holds, and how they connect. Ten minutes here prevents the most common early mistake, which is modelling your products wrongly and having to redo them.

The free plan is the right place to start: ৳0, 2 team members, 25 warranties, 25 claims, 25 service requests, 25 invoices and 25 sales a month, analytics, and inventory and expense management. That is a real evaluation budget and a deliberately small one.

Day one to two: build the catalogue

Add the products you shortlisted. For each, you want the identifying details and whatever your staff will actually search by at the counter — the name they use, not the name on the manufacturer's datasheet.

This is the step that determines how fast registration is later. A catalogue that matches how your team thinks about products makes registration a four-second lookup. One that matches how your supplier's spreadsheet is organised makes it a hunt.

Resist adding the long tail now. You can add products as you sell them; you cannot easily undo a catalogue structure you got wrong across three hundred lines.

Day two to three: configure warranty terms

Attach coverage to products or product lines: the period, the inclusions, the exclusions.

Two things to get right at this point, because they are painful to retrofit:

Be specific about exclusions. "Normal wear and tear" is not a rule anyone can apply at a counter. "Consumable parts listed in section 3, and any damage from liquid ingress" is. Every vague clause becomes an argument between a customer and a member of staff who has no backup.

Understand that terms are bound at registration. The version in force when a unit is registered is the version that applies to that unit for its whole life. That is what makes a claim assessable years later — the reviewer reads the contract that actually applied, not whatever is current. It also means getting terms right before you register volume is worth an extra hour today.

Day three to four: register live sales, not history

Switch on at the till. From this day, every qualifying sale gets registered: product from the catalogue, serial captured, customer attached, confirmed. The customer receives a QR-verifiable digital certificate immediately.

This is the part that takes real effort, and it is worth saying clearly: the hard problem is not the software, it is the habit. A registration takes seconds, but it takes those seconds during a transaction, when there is a queue and the customer wants to leave.

What helps:

• Have one person check the daily registration count against the daily sales count. A gap that nobody looks at grows.
• Make the step short enough that nobody has to choose between doing it properly and serving the next customer.
• Tell staff what it is for, with a real example. "The claim we approved twice last year" is more persuasive than a policy.

What does not help: a launch email, and then nothing.

If you are moving off a spreadsheet, the migration path from spreadsheet to warranty system covers the data side in more detail. The behaviour side is the same either way.

Day four to five: invite your team

Add the people who register warranties and the people who approve claims, with the roles that match what they actually do. A counter assistant does not need the permissions of a manager who sees warranty cost reporting.

One plan constraint to plan around: the free plan has two team member seats. If more people need to log in — which is true for most real shops — you need Starter at ৳1,230 a month ($9.99) for up to 5 team members, or Professional at ৳2,460 a month ($19.99) for up to 15. This is usually the first limit a business hits, long before the warranty count.

Starter and Professional both start with a 14-day free trial as soon as you choose them, with no card needed. We issue an invoice when the trial starts, due when it ends, which you can pay by bank transfer, bKash, Nagad, Rocket or cash; the plan continues once the payment is verified. Moving down to Free takes effect immediately.

Day five: run one test claim, end to end

Do not skip this, and do not do it in your head.

Register a real product to a test customer. Open a claim against that serial. Watch what the platform returns: coverage checked against the registered terms, the expiry arithmetic, duplicate detection against the unit's history, serial validation, ownership, and the terms clause that applies. Every one of those is a deterministic lookup against a record — nothing is predicted or scored, so every flag can be traced back to the document that produced it.

Then take it all the way: review it, record a decision, track the service work against the same record, notify the customer, close it, and read the audit trail that remains. How Warranlytics handles a claim end to end walks through the same sequence in detail.

Now run the awkward version. Claim against an expired warranty. Claim twice against the same serial. Claim against a serial that does not exist. You want to see the exception path before a real customer is standing in front of it, and you want your staff to have seen it too.

Note that a test claim uses one of the free plan's 25 claims. Budget for that.

Day five onward: decide what to migrate

Only now, with live capture working, is the archive question worth answering.

Data	Worth migrating?	Why
Warranties still in coverage	Yes	These will generate claims; without them, claims arrive against units the system does not know
Warranties expiring within 3 months	Probably not	High effort, short remaining life
Expired warranties	No	Keep the old records for reference; do not clean them
Customer records	Yes, for live warranties	A warranty with no owner cannot be transferred or notified
Historical claims	Only against live warranties	Duplicate detection needs them; older ones add noise
Product catalogue	Already done	You built the useful subset on day two

The realistic cost here is data cleanup, not import. Duplicate customers, serials recorded three different ways, sale dates in two formats, warranties with no serial at all. That is human work and it takes longer than anyone expects, which is precisely why it belongs after live capture is running rather than before.

Start with the narrowest useful set: units still in coverage, with a serial, with an owner. Everything else is optional.

When do you need to upgrade?

Four thresholds, in the order most businesses meet them:

1. A third person needs to log in. Free has 2 seats. → Starter, with 5.
2. You pass 25 warranties, 25 claims, or 25 invoices in a month. → Starter, at 300 each.
3. You need CSV and Excel export, or SMS for invoices and sales. → Starter.
4. You need unlimited volume, custom branding, API access, or more than 5 seats. → Professional, at up to 15 team members.

If you are adding Warranlytics alongside systems you are keeping, automating claims without replacing everything is the relevant read. And once a few months of registrations have accumulated, the warranty data you are not collecting covers what to look at first.

---

A workable first week is narrow on purpose: a small catalogue, correct terms, live registration, one tested claim. Migration can wait; capture cannot. The retailer-specific version of this is in registering warranties at the point of sale, and current limits are on the pricing page.