About Bill Hartzer

One person, and the same person throughout

I am Bill Hartzer. I take the first call, review the account export, write the opinion and sit for the deposition. There is no associate doing the analysis under my name and no research department behind me.

That is a limit as much as a feature: it caps how many matters I can carry at once. What it buys you is that the person who formed the opinion is the person who has to defend it, with no seam between the analysis and the testimony where a hand-off can go wrong.

Where the edge of my expertise sits, I say so on this page rather than discovering it in a deposition. The fastest way to lose an opinion is to have signed it outside the area you can actually support.

Where the experience comes from

I have been working in search marketing since 1996 — which is to say I was working on how businesses appear in search results before paid search, as an auction-priced advertising channel, existed at all. I watched placement go from something bought on a flat monthly basis to something priced in a live auction milliseconds before it renders. That is a large part of why the field has no settled professional norms: the mechanism changed faster than any body could write standards for it.

The work moved with the channel. Early questions were about visibility and how a page got found. When keyword-priced advertising arrived they became match types, bids and quality scoring, then conversion tracking, then attribution modeling, and most recently automated bidding and campaign types where the advertiser no longer picks the keyword at all. Knowing which era's mechanics governed the conduct in dispute is half of reading a record correctly.

I founded Hartzer Consulting and I founded DNAccess — the first search marketing work, the second domain-side work including acquisition, recovery and the investigation that goes with it. The expert witness practice grew out of both, because attorneys kept arriving with a question the operating work answers directly: what does the record actually show?

What I have actually done inside an advertising account

The credibility of an examination rests on having done the thing examined. Concretely:

• Building and running search, shopping and display campaigns on Google Ads and Microsoft Advertising, including the budget, bid, match-type and negative-keyword decisions that later become the subject of a dispute.
• Installing and repairing conversion tracking — the tag, the conversion action definition, the counting setting, the attribution model — and then explaining why the reported number moved when nothing about the business did.
• Publishing and rolling back tag manager container versions, frequently the only surviving record of when a tag was changed, broken or removed.
• Reconciling platform billing against agency invoices, which is the whole of some disputes and the beginning of most.
• Exporting change history from the interface and pulling the same period through the API, and seeing the two disagree.
• Moving accounts between manager accounts, where ownership and access stop being abstract.
• Reading search terms reports closely enough to know what is missing from them.

What operating an account teaches that reading about it does not

Google's own developer documentation states that a change event “may not include every row from the Change History in the web client.” An examiner who pulls the programmatic record and one who exports the interface record will produce different sets from the same account over the same period. If neither says so, the discrepancy becomes the cross-examination instead of a footnote.

Search Ads 360 displays timestamps in the viewer's browser time zone and converts them to GMT on download. Compare a downloaded change history against a screenshot of the same screen and the same event carries two different times — an artifact of retrieval, not an inconsistency in the record, and it belongs in the report rather than in a surprise question.

That download also captures only the columns visible in the table at the time, silently omitting fields that were present in the record. A production can be complete as produced and incomplete as evidence.

The adjacent work, and where it is load-bearing here

Alongside paid search I work in technical SEO, website and domain migrations, domain names and domain recovery, and internet investigations. Most of that is a different subject. Two pieces of it carry weight here.

The first is investigation. Establishing who controlled a domain, a hosting account or a set of landing pages on a given date is the same discipline as establishing who controlled an advertising account, and affiliate and lead-generation matters usually turn on exactly that.

The second is migration. Moving a site or an account exposes what breaks silently: a conversion tag that never came across, a redirect chain that strips tracking parameters, an analytics property whose retention setting was left where somebody found it. Those are the same failures that appear in litigation, seen from the other end and usually years earlier.

What I am not

I am not an attorney. Nothing I write or say is legal advice, and whether a set of facts is actionable is a question for counsel. My side of that line is what the account record shows and what it does not.

I am not a forensic accountant or an economist. Where a matter needs a lost-profits figure, that figure belongs to someone with that qualification. My contribution sits underneath it: what the account did, what was spent, what portion of the spend was recoverable, and whether two comparison periods measure the same thing at all.

I am not the right witness on admissibility procedure as a general subject. I can describe how a record was retrieved and why an export a party generated is not self-proving; the rules governing whether it comes in are counsel's territory.

How I frame an opinion, and why it is framed that way

There is no published, citable professional standard of care for paid-search account management. No licensing board, no accrediting body, no professional association with practice standards and disciplinary authority. That absence is a finding rather than a gap to paper over, and it decides how an opinion here has to be built.

An opinion that says “no competent manager would have done this” rests on a standard that cannot be produced when it is asked for. An opinion that says “the conversion tag was absent between these dates, here is the tag record, here is the spend that ran, and here is the arithmetic” rests on the record. Since the amendment to Federal Rule of Evidence 702 effective December 1, 2023, the proponent must show that an opinion reflects a reliable application of method to the facts of the case, and only the second framing carries that weight.

So I work as a record examiner rather than a declarer of norms. Where an argument about reasonable practice is available at all, it is built from the contract and scope of work first and the platform's own published documentation second — documents either side can check.

What I decline

I do not take an engagement where the conclusion is specified in advance, and I do not accept a fee contingent on the outcome. I do not sign a report I could not defend line by line, which means the section naming what the record cannot establish is drafted before the findings are.

And I say no when the account data cannot reach the theory. Some paid-search disputes are contract disputes wearing a data costume: the argument is about what was promised or what somebody intended, and no examination of the account settles either. Saying that on the first call costs me the engagement and saves a client the cost of an examination that was never going to answer the question.