How to read a figure
Each figure is tagged with what kind of fact it is, its source and its date. Estimates, inferences and our own calculations are always labelled as such.
- Observed
- Seen directly, such as a search result, a map listing or an ad.
- Official
- Published by a government source.
- Estimate
- A third-party estimate, such as search volume.
- Inferred
- Our reading of indirect signs, such as a service line found on a website.
- Model
- Our own calculation from published figures.
Areas
Areas are the ones the Australian Bureau of Statistics uses (SA3, about the size of a council; SA2, about a suburb), named the way people say them. Codes appear only on pages like this one. Where a publisher still uses the 2016 boundaries, we share its figures out to the current ones by population and say so.
Matching providers
The NDIS provider register is our starting list. Search results, map listings, ads and profiles are matched to it by ABN, then website, then phone number (compared as a one-way code, never stored as a number), then name in the same area. Matches we are unsure of are checked by an analyst before they count.
A business we can’t match to the register is shown as “Not found in the NDIS provider register”. That describes our matching, not the business: it may be registered under another name, or not need registration for the services it offers.
Who counts as a competitor
A provider competes in an area for a service when it offers that service (by its registration, or by signs on its website and Google listing) and it is present there: a service area listed on the register, a Google listing with its postcode in the area, or a place in the area’s own search results for the service. A brand and the group it belongs to count once.
Across several areas, a provider serving more than one is counted once, and its shares of the search results are worked out over all the areas’ neighbourhoods together.
Providers not on the register
Some businesses serving participants appear on Google but not on the register. We add one only when the evidence agrees: a sign it offers these services (its Google category or its own website), signs it is a business rather than an individual, at least two independent sources (or one and an ABN on its website), and a presence in the area. Borderline cases are checked by an analyst.
Only registered providers may offer Supported Independent Living, Specialist Disability Accommodation, plan management and behaviour support. A business offering one of these that we can’t find on the register isn’t counted for it: it is usually a registered provider we haven’t matched yet, and an analyst checks it.
Takeovers, closures and duplicates
A website saying a business is now part of another, an ABN cancelled while the business still trades, or Google marking every listing permanently closed are signs, not facts. None changes what you see until an analyst has confirmed it.
Coverage and freshness
Prodja covers South Australia today; other states follow. Each source is refreshed on its own schedule (see Sources below), and every figure shows the date it was collected, so you can see how current it is. Figures from a quarterly release describe the quarter the publisher reports, which can be a few months before today.
Measures
Your market
- Participants
The number of participants living in the area, as the agency publishes it.
- Participant growth
How much the number of participants in the area changed since the last quarter and the same quarter last year.
- Participants per provider
How many participants there are for each active provider in the service district.
- Market concentration
The share of payments that goes to the ten largest providers in the service district.
- NDIS interest (state)
How much people in the state are searching Google for general information about the NDIS this month, against the past 12 months.
- Plan-managed participants
Participants whose NDIS plan is managed by a plan manager, in the service district or the state, and their share of all participants there.
- Plan-managed participants (estimate)
An estimate of the participants in an area whose NDIS plan is managed by a plan manager.
Visibility on Google
- Visibility share
The share of Google’s first ten ordinary results a provider holds when people in the area search for a service.
- Organic position
Where a provider ranks on average in Google’s first ten ordinary results when people in the area search, and in how many neighbourhoods it appears there.
- Local pack share
The share of places in Google’s map pack (the map and the listings under it) a provider holds for searches in the area.
- AI Overview citation share
The share of the sources cited in Google’s AI Overviews that belong to a provider, where Google shows an AI Overview.
- Directory dominance
How much of the first page of results and the map pack goes to directories rather than to providers’ own sites.
- Platform share
Each directory’s and matching platform’s share of the first page of results and the map pack in the area.
- Local pack position
A provider’s best place in Google’s map pack, per neighbourhood and search.
- Google profile completeness
How many of four basics a provider’s Google Business Profile shows, compared with the businesses in the area’s map packs.
Advertising
- Google ads active
Whether a provider had Google ads running in the month, with when they were first and last seen.
Competitors
- New reviews per 30 days
How quickly a provider is collecting Google reviews.
- Compliance actions
Compliance actions the regulator has published against a provider.
Market moves
- New entrants
Providers that started serving the area in the month.
- Exits
Providers that stopped serving the area in the month.
Opportunities
- Search demand
About how many Google searches a month there are for a service in the area’s search market.
- Providers per 1,000 participants
How many registered providers offer a type of service in the area for every 1,000 participants.
Sources
| Source | What we use | Refreshed |
|---|---|---|
| NDIS provider register | Registered providers, their registration groups, and the areas they list as service areas. Service areas are used for coverage and never shown as addresses. | Weekly |
| Compliance actions published by the NDIS Quality and Safeguards Commission | Compliance actions the regulator publishes. | Weekly |
| NDIA, participants by area (quarterly) | Participants by area, and the agency’s figures on active providers, market concentration and projections. | Quarterly |
| Google search results and map packs, checked across the area | Google results, map packs and AI Overviews for each service’s searches, from every neighbourhood. | Monthly |
| Google Maps listings | Google Maps listings for each service across the area, to find providers and their locations. | Quarterly |
| Google Business Profiles | Google Business Profiles: category, photos, hours, website, rating and review count. No review text or reviewer names. | Monthly |
| Google Ads Transparency Center | Whether providers run Google ads, from Google’s Ads Transparency Center. | Monthly |
| Google Ads keyword volumes (monthly searches) | Google’s monthly search volume estimates for each service’s searches. | Monthly |
| Providers found on Google and websites, checked by Prodja | Providers not on the register that we find on Google and confirm on their own websites, read politely. | Monthly |
| ABN Lookup (Australian Business Register) | A business’s ABN, its status and its registered names, to confirm who a business is and whether it still trades. | Weekly |
| ACNC Charity Register | Whether a provider is a registered charity, and the financial size it reports. | Quarterly |
| Online directories of providers | Whether a provider is listed in the main online directories, and at which level. Not what the listing says. | Quarterly |
| Providers’ own websites | Signs on a provider’s own website, such as the services it describes or a notice that it has closed or joined another business. | Monthly |
| Australian Bureau of Statistics, statistical areas and postcodes (ASGS Edition 3) | The areas and their boundaries, population, and which postcodes fall in each. | With each ABS edition |
Google data is collected through a licensed data provider. Websites and directories are read politely: we follow each site’s robots.txt, read a few pages at a time, and identify ourselves as ProdjaBot (prodja.com.au/bot).
Limits of the data
- Search results are collected once a month, from a point in each neighbourhood. What a person sees can differ with where they are, what they searched before and when. Shares and positions describe these samples, not every search.
- Search volumes are Google’s estimates, rounded, and not published for every place or search. They show demand, not the number of participants.
- Official figures are published some months after the period they describe, and publishers hide small counts (for example “<11”) to protect privacy. We show them as they are published.
- The NDIS provider register shows registration, not how much a provider is active, and may lag behind changes.
- Matching records to providers can be wrong, and websites and listings say what businesses choose to publish. Tell us when something looks wrong (see Corrections).
Attribution and licences
- © National Disability Insurance Agency. NDIS data is licensed under CC BY 4.0.
- Provider registration data: NDIS Quality and Safeguards Commission, NDIS Provider Register.
- Areas: based on Australian Bureau of Statistics data (Australian Statistical Geography Standard, Edition 3), licensed under CC BY 4.0.
- Google data is collected through a licensed data provider. Prodja shows its own analysis of it, not the underlying data, and isn’t affiliated with Google.
- Government sources are named as text; we don’t use their logos, and nothing in Prodja suggests their endorsement.
What we never collect
- Anything about participants, and no personal contact details.
- Reviewers’ names or what reviews say: only counts and ratings.
- Social media posts or media: at most, activity counts.
- Individual workers or their profiles on worker platforms.
- Addresses from the register: they can be where services are delivered, even people’s homes, so we use them only to count coverage by area.
- Phone numbers as numbers: only a one-way code, used for matching.
- Data whose terms don’t allow commercial use.
Corrections
If something Prodja shows about a business is wrong, such as a match to the wrong business, a service it doesn’t offer or a closure that didn’t happen, email privacy@prodja.com.au with what’s wrong and how we can check it. An analyst looks into every report and replies within 30 days; a correction shows as soon as it’s confirmed.