Doing It Yourself vs Hiring a Locate
The usual version of this argument is that free tools are shallow and paid ones are deep, and it is unfalsifiable, which is why every vendor makes it. There is a better version, and since 1 August 2026 it has a date on it. This page is for anyone deciding whether to keep running a search themselves or hand it over – creditors, attorneys, process servers, landlords and insurers – and People Locator Skip Tracing takes the ones that have stalled. California’s Delete Act splits the locating world in two by name: its definition of a data broker excludes entities to the extent they are covered by the Fair Credit Reporting Act, by the Gramm-Leach-Bliley Act, by the insurance privacy statute, or by the health-information exemption. Everything else – the consumer-facing aggregate sites a self-serve search actually runs on – sits inside the definition, and from 1 August 2026 those businesses have to check the state’s deletion platform at least every forty-five days and erase what a single verified request tells them to erase. So the question is no longer whose database is better. It is which tier your answer came from, and whether it can name the record it came out of when somebody asks. This page is about that difference. This is a records-research firm, and everything below is general information rather than legal advice.
A Statute Draws the Line, and It Named the Tiers
Read the definition and the split stops being a marketing claim.
California’s Delete Act – Senate Bill 362, which amended this section by Chapter 709 of the 2023 statutes with effect from 1 January 2024 – has to say what a data broker is before it can regulate one. Civil Code section 1798.99.80(c) does it in two moves. A data broker is a business that knowingly collects and sells to third parties the personal information of a consumer with whom the business does not have a direct relationship. Then it excludes four things, each of them phrased as “an entity to the extent that it is covered by”:
- the federal Fair Credit Reporting Act;
- the Gramm-Leach-Bliley Act and its implementing regulations;
- the Insurance Information and Privacy Protection Act;
- an entity, or a business associate of a covered entity, to the extent its processing is exempt under section 1798.146, the health-information exemption.
Look at what that list is. It is the set of federal and state regimes that condition access on the requester’s status and purpose. The legislature did not exclude them because their data is better. It excluded them because they are already regulated somewhere else, on terms that reach who may receive information and for what.
The half that is not excluded is the half a self-serve search runs on
Everything outside those four exclusions – the aggregate sites, the reverse-lookup pages, the free-then-paywalled people finders – is a business selling personal information about consumers it has no relationship with. That is the definition, not a characterisation. And it is why the free tier feels the way it feels: it is assembled from whatever can be bought and scraped, held by a business the subject has never dealt with, and governed by no permissible-purpose architecture at all.
The point is not that this tier is worthless. It is that its position in the regulatory map explains everything people complain about it, and now predicts what happens to it next.
Since 1 August 2026, the Free Tier Has to Delete on Request
This is not a prediction. It is an obligation with a commencement date.
The Delete Act required the California Privacy Protection Agency to build a single mechanism through which a consumer could reach every registered broker at once. The agency’s own data broker page states the resulting duty plainly: under the Delete Act, beginning 1 August 2026, data brokers must access the accessible deletion mechanism at least once every forty-five days and process consumer deletion requests, subject to limited exceptions, under Civil Code section 1798.99.86(c). The agency describes the mechanism as allowing a consumer, through a single verifiable request, to direct every data broker holding personal information about them. Its instruction to brokers is equally concrete: starting 1 August you have forty-five days to access the platform and process your first batch.
Registration is separate and annual. A business that brokers the data of California residents registers each year, and the agency notes that a broker failing to register by 31 January may be liable for administrative fines and costs in an action or investigation it brings, under Civil Code section 1798.99.82(d).
What that changes about a DIY search, concretely
It converts an old complaint into a mechanism. “The free sites are out of date” used to be a grumble about data hygiene. It is now a description of a legal process running on a forty-five-day clock, in one direction, initiated by the very people a searcher is looking for. Every consumer who files one request removes themselves from every registered broker simultaneously, and keeps removing themselves as new data arrives.
Two honest qualifications, because this page is not going to overclaim a statute it just read. The Delete Act is Californian; other states have their own broker-registration regimes and we have not enumerated them, so no claim here is quantified over the other forty-nine. And the exclusions in 1798.99.80(c) are drawn around regulatory coverage, which is a question about a particular entity’s particular processing – the page describes how the statute draws its line, and makes no assertion about any individual vendor’s status, including our own suppliers’.
What survives both qualifications is the direction of travel. The tier a self-serve search depends on is the tier the law is now emptying, on a schedule, at the subject’s own request. Building a locate on it is building on the one part of the map that is legally obliged to shrink.
The Short Version
Ask which tier the answer came from, and whether it can name its own source. The self-serve tier is aggregate data sold by businesses that have no relationship with the person it describes – which is the actual statutory definition of a data broker in California, at Civil Code 1798.99.80(c), whose four exclusions are drawn around the FCRA, the Gramm-Leach-Bliley Act, the insurance privacy statute and the health-information exemption. Since 1 August 2026 that tier has had to check California’s deletion platform at least every forty-five days and erase what a single verified consumer request tells it to erase. The professional tier is not defined by having a bigger database; it is defined by working under regimes that condition access on the requester’s status and purpose, and by producing a finding that names the record and the date behind it. That second half is what a self-serve result structurally cannot do, and it is what an affidavit of diligent search, a declaration or a motion actually needs. This firm is not a consumer reporting agency; a locate is not a consumer report; and a tenancy, a hire, an extension of credit or an insurance rating cannot lawfully be decided on one. We publish no accuracy percentage, ours or anyone else’s. General information, not legal advice.
The Thing a Self-Serve Result Cannot Carry
Not depth. Provenance – and the namesake problem it exists to solve.
Suppose the free tier were perfectly current. There would still be a difference, and it is the one that matters where the address is going.
An aggregate has no custodian
A people-search entry is a claim with no author. It cannot tell you which record produced it, when that record was created, or which office holds it, because the aggregation step discarded that information on the way in. So it cannot be checked, it cannot be re-pulled, and it cannot be put in front of anybody whose job is to test it. A recorded instrument has a book and page. A court file has a case number and a clerk. A registration has a filing date and a registrant’s signature. The whole professional method is the refusal to strip that layer off.
The namesake problem is an adjudication, not a lookup
Free tools mix people up not because they are careless but because they have no adjudicating step: they cluster records by name similarity and present the cluster. Separating a person from a namesake is a decision made against corroborating evidence – a middle initial in one filing, an age band in another, a shared address history, a spouse’s name on a deed, a matching date on a court docket – and somebody has to make it and be accountable for it. A tool that shows you six Michael Alvarezes has not done that work; it has handed it to you unlabelled. Our note on how to verify a skip tracing report sets out what a made and labelled decision looks like on paper.
Where the address is going decides how it has to be built
If it is going into a Christmas card, none of this matters. If it is going into an affidavit of diligent search, a declaration in support of alternative service, a motion, a garnishment or an estate accounting, then somebody adverse will eventually ask how it was obtained – and the honest answers are either “the recorder’s index for that county, on this date” or “a website”. Only one of those survives the question. That is also the practical reason the professional tier exists at all: not to know more, but to be answerable.
And one limit that runs in the other direction
The federal permissible-purpose architecture at 15 U.S.C. 1681b(a) closes its own list with the words “and no other”, and there is no purpose on it called locating a person. So a professional locate is not a consumer report and must never be used as one – not for a tenancy, a hire, credit or insurance. If the underlying decision is one of those, no tier of locating is the right product, and the answer is an FCRA-regulated screening provider. That boundary is not a weakness of the professional route; it is the same status-and-purpose logic that draws every other line on this page.
Two things neither tier gets from us
The first is deception. Pretexting a financial institution is a federal offence in its own right, and beyond that, a finding obtained by deception is worthless the moment its origin is examined – which is the same provenance argument this whole section is making, seen from the other end.
The second is a class of request. We turn down work whose object is to find a person who moved in order to be safe – matters touching domestic violence, sexual assault, stalking, trafficking, protective orders and no-contact orders – and we turn it down when the shape of a request says so even where the stated reason is ordinary, such as a requester who will not identify the interest they hold or an unexplained hurry about somebody who has just moved without leaving word. Confidential address programmes exist for this in most states, and an advocate can reach one faster than any search would. That decline stands on both tiers, and it is worth saying on a page comparing them: doing it yourself does not make it a different question.
Six Points Where Self-Serve Stops
Structural limits, not effort. More hours on the same tier does not clear any of them.
The Subject Filed a Deletion Request
One verified request reaches every registered broker, and from 1 August 2026 they process on a forty-five-day cycle. There is nothing left to search harder.
Six Candidates, No Adjudicating Step
An aggregate clusters by name similarity and stops. Somebody still has to decide, against evidence, which one is the person.
The Result Has to Go in a Filing
An entry with no custodian, no date and no office behind it cannot answer the question a court asks about how it was obtained.
The Trail Crosses a County or a State Line
Recorded and filed records live in named offices with their own indexes and date floors. A national aggregate flattens all of that away.
The Last Real Address Is Twenty Years Old
Cold files are worked backwards through recorded instruments and court filings, which is a records skill rather than a query.
The Purpose Needs to Be Documented
A self-serve search records nothing about why it was run. If that ever matters, it matters retrospectively and there is nothing to produce.
Watch: Which Tier Answered You
Why the split is regulatory rather than technical.
Watch Overview
What Happens When You Hand It Over
Bring the dead ends. They are evidence, and they shorten the work.
Send the Failed Search, Not Just the Name
The addresses that came back, the ones you ruled out, the candidates you could not separate. Each is a data point about where the trail has already been walked, and none of it has to be repeated.
State the Purpose in a Sentence
Written down before anything is opened, because lawful sources condition access on it and because a purpose recorded later reads exactly like one recorded later.
Resolve the Person Before the Address
The namesake question is settled against corroborating records first. An address attached to the wrong person is worse than no address, because it gets acted on.
Get Back a Finding With Its Source Attached
The record, the office and the date behind every line, plus a plain account of what could not be established. That is the part you could not have produced yourself, and it is the part that has to survive being questioned.
Who Crosses This Line and Why
Usually at the point where the answer has to be defensible rather than merely present.
Litigators Filing on Service
Where a declaration has to describe how the address was obtained
Process Servers With Three Bad Attempts
Needing a current door rather than another guess at the old one
Creditors Holding a Cold Judgment
Where the last good address predates most of the online record
Estate and Trust Administrators
Tracing heirs whose names have changed and whose files must balance
Relatives Looking for Relatives
Where a free-site cluster produced six candidates and no answer
Anyone Who Already Tried
And wants the hours already spent to count for something
What We Will Not Claim in This Comparison
No accuracy percentage appears anywhere on this page. The pages ranking for this question publish two of them, in ranges, with no source behind either, and repeating an unsourced figure to win a comparison is precisely the practice this page is arguing against. We do not name commercial databases as trophies, we do not publish a per-record price to be compared against a scraper’s, and we do not tell you that self-serve searching is useless – for an easy, current, uncontested name it often is not. What we will tell you is where it structurally stops, why a California statute now draws that line with dates on it, and what a finding has to carry to be worth anything after somebody starts asking questions about it. Every legal fact here was read in the enacted text or on the agency’s own page, and where something could not be settled – the position in the other forty-nine states, the coverage status of any particular vendor – the page says so rather than filling the gap.
DIY and Professional Questions
Can I just use a free people-search site myself?
For an easy, current, uncontested name, often yes, and nothing here says otherwise. What is worth knowing is what that tier is in law rather than in marketing. California’s Delete Act defines a data broker at Civil Code 1798.99.80(c) as a business that knowingly collects and sells to third parties the personal information of a consumer with whom it has no direct relationship, and then excludes entities to the extent they are covered by the Fair Credit Reporting Act, the Gramm-Leach-Bliley Act, the Insurance Information and Privacy Protection Act, or the health-information exemption at section 1798.146. The consumer-facing aggregate sites are inside that definition. That is the tier a self-serve search runs on, and it is a different regulatory animal from the sources that condition access on the requester’s status and purpose.
What actually changed on 1 August 2026?
The obligation to delete became operative. The California Privacy Protection Agency states that under the Delete Act, beginning 1 August 2026, data brokers must access the accessible deletion mechanism at least once every forty-five days and process consumer deletion requests, subject to limited exceptions, under Civil Code section 1798.99.86(c). The agency describes that mechanism as letting a consumer, through a single verifiable request, direct every data broker holding personal information about them, and its guidance to brokers is that from 1 August they have forty-five days to access the platform and process their first batch. Registration is separate and annual, and a broker that fails to register by 31 January may face administrative fines and costs under section 1798.99.82(d). This is a Californian statute; we have not enumerated what the other states require.
Does that mean professional data is exempt from deletion?
That is not what the page says and it is not what we would claim. The statute’s exclusions are written as “an entity to the extent that it is covered by” a named regime, which is a question about a particular entity’s particular processing rather than a blanket status, and we make no assertion about any individual vendor – including our own suppliers. What is verifiable is the shape of the line: the legislature drew its exclusions around the FCRA, the GLBA, the insurance privacy statute and the health exemption, which are the regimes that already condition access on who is asking and why. The practical reading is that the split between tiers is regulatory rather than technical, which is the whole argument of this page.
Why do free sites mix up different people?
Because there is no adjudicating step in the product. Aggregates cluster records by name similarity and present the cluster; deciding which member of the cluster is the actual person is work the tool has handed back to you, unlabelled. A professional locate makes that decision explicitly and against corroborating evidence – a middle initial appearing in one filing, an age band in another, an overlapping address history, a spouse named on a recorded deed, a matching date on a court docket – and then says in the report which evidence carried the decision. An address attached to the wrong person is worse than no address, because unlike a blank it gets acted on.
Will a DIY result hold up for a court or a creditor?
The question a court asks is not whether the address is right but how it was obtained, and that is where a self-serve result has nothing to say. A people-search entry has no custodian, no filing date and no office behind it, because the aggregation step discarded all three. A recorded instrument has a book and page. A court file has a case number and a clerk. A business registration has a filing date and a registrant. If the address is going into an affidavit of diligent search, a declaration supporting alternative service, a motion, a garnishment or an estate accounting, it needs to be the second kind, and the deliverable needs to show it.
Is a professional service just paying for a bigger database?
No, and framing it that way is how the comparison goes wrong. The difference is status, purpose and provenance. Status and purpose because the regimes named in the exclusions above – the FCRA and the GLBA in particular – condition access on who is asking and for what, so the relevant question is which category a requester occupies rather than which product they can find. Provenance because the deliverable either names the record behind each line or it does not. We publish no accuracy percentage to support any of this, in either direction, because the figures circulating on this question have no source attached to them.
I have already spent hours on it. Is that wasted?
It is the opposite of wasted, and it is worth sending. The addresses that came back, the ones you ruled out and why, the candidates you could not separate, the relatives you already identified – each of those is a data point about where the trail has been walked, and none of it needs repeating. Handing over a documented dead end is usually the cheapest possible start, because the expensive part of a cold file is establishing what is already excluded. Bring the failures as well as the leads.
What can a professional locate never be used for?
A screening decision. The Fair Credit Reporting Act’s permissible-purpose list at 15 U.S.C. 1681b(a) closes itself with the words “and no other”, and there is no purpose on it called locating a person – Congress regulated the decisions rather than the act of finding somebody. So a locate is not a consumer report, this firm is not a consumer reporting agency, and nothing we deliver may be used to decide a tenancy, a hire, an extension of credit, or the writing or rating of insurance. If that is the underlying decision, the right product is a compliant consumer report from a consumer reporting agency, and we will say so at intake instead of selling you something that would be misused.
Hand Over the Search That Stalled
Send the name, whatever identifiers you already hold, every address that has come back so far including the ones you have ruled out, and the lawful reason the search is being run. What comes back is a resolved person rather than a cluster, an address with the record, office and date behind it, and a written note of what could not be established. If you want the ground underneath this first, what skip tracing is covers the definitions, how skip tracing works covers the mechanics, how much skip tracing costs covers the commercial side, and the skip tracing services hub covers the rest. Workable requests usually get a first read back within 24 hours. Contact us when the free tier runs out.
Start Your Request →
