Your model informs their decision, and if it's wrong, you're the one they blame
Clients hand you the data and act on what you tell them. Get the model wrong, mix up records in a pipeline, or let an exported dataset sit somewhere it shouldn't, and when it costs them money or exposes their customers, the analyst is the first call. Professional indemnity and cyber cover, built for data work.
Where a data analyst actually gets exposed
Your risk isn't a system going down. It's a number being wrong, or a dataset going somewhere it shouldn't. If a client can argue your analysis or your handling of their data fell short and it cost them, that's a claim, and standard business cover won't answer it.
The forecast that missed by a mile
You build a demand forecast that a retail client uses to plan stock. A flawed assumption in the model leads them to over-order significantly, tying up capital in stock that doesn't move.
They argue a competent analyst would have caught the flaw, and bill you for the cost of the excess inventory.
The pipeline that mixed up the records
A data pipeline you built merges and de-duplicates customer records feeding a client's billing system. A logic error incorrectly matches accounts, and hundreds of customers are billed against the wrong record.
The client wears the refunds and complaints, and passes the remediation cost back to you.
The export that went out unencrypted
You export a full customer dataset to build a model offline, and it's stored on a laptop that's later lost or accessed without authorisation. Under the Privacy Act, your client has to notify everyone in the file.
They say the data never should have left their systems that way, and look to you for the response costs.
Who's going to ask you for it
It's how you get access to the data
For analysts and data consultants, cover is often the price of admission to the client's systems and datasets:
- Enterprise data contracts commonly require $1M–$5M professional indemnity before you're granted access to production data.
- Prime contractors and agencies subcontracting you flow their own insurance requirements down before you can start.
- Government data panels set minimum PI and cyber limits, often higher given the sensitivity of the data involved.
The exclusion analysts assume is covered
Two things catch analysts out. First, garbage in, garbage out: relying reasonably on data the client supplied is fine, but a claim that you failed to sanity-check or QA an obviously flawed dataset before basing recommendations on it can still land on you.
Second, "I don't touch systems, so no cyber risk" is a common misconception. Holding an exported dataset is enough to trigger cyber exposure, even without ongoing system access.
Common questions
I just build reports and dashboards, where's my risk?
The risk is in the decision the report drives. If a client acts on a flawed number and it costs them, that's a professional indemnity claim regardless of how simple the deliverable looked.
Isn't this just professional indemnity?
PI covers the analysis itself, but handling client datasets, even briefly, even offline, carries real cyber exposure. ICT cover combines professional indemnity with cyber so there isn't a gap between policies.
What limit should I get?
Take it from the client's data access agreement or the prime contractor's flow-down requirement, it will usually name a minimum. If you're unsure, we'll read it with you before you buy.
Does it cover analysis I did before I was insured?
Professional indemnity is claims-made and usually responds to claims first made while you're insured, provided the issue wasn't already known to you. It's best to have cover in place before a dispute surfaces.
A client needs a certificate of currency before granting data access, how fast?
Once bound, a certificate is quick to issue. If access is waiting on it, tell us and we'll prioritise it.