A law firm ran five months with no conversion tracking. Here is what fixing it revealed.
Case study · Personal injury law, measurement
Five months of zero conversions. The phone was ringing the whole time.
A personal injury firm’s analytics recorded no phone calls and no form fills between October and February. Traffic never stopped: the site took between 774 and 1,316 sessions every one of those months. The tracking had simply gone dark, and nobody noticed.
How to read these numbers. The window is the 90 days ending 21 May 2026, compared with the 90 days before it (22 November 2025 to 20 February 2026). Same property, same tool, same length of window on both sides. Where it matters, brand searches are separated from non-brand ones, because a rise in people typing the company name measures recognition, not new demand. Every figure below is pulled from the client’s own Google Search Console or GA4 property. Client names are withheld; the industry, the numbers and the method are exactly as they happened. Full methodology.
What was actually wrong
This is the most common expensive problem in small-business marketing and it almost never appears in a case study, because it makes everybody look bad: the agency, the previous agency, and the client who signed off on reports full of zeros.
For five consecutive months the firm’s analytics recorded no conversions at all. Not a low number, zero. Meanwhile sessions ran between 774 and 1,316 a month, ads were running, and the phone was ringing. Every decision made in that period, about budget, about channels, about which pages deserved investment, was made on a number that was not measuring anything.
The damage from broken tracking is not the missing data. It is that the missing data looks exactly like failure, so the natural response is to cut the channels that appear to produce nothing, which usually means cutting the ones that were quietly working.
How the work was done
Four moves, in the order they were made. The order matters more than the list: measurement first, then the structural repairs, then the pages.
Found the gap by looking at the shape of the data, not the total
A month at zero can be a bad month. Five consecutive months at zero, against steady traffic, is not a business result, it is a measurement failure. The tell is the flat line: real demand is noisy, broken tracking is perfectly smooth.
Rebuilt the events that represent actual business outcomes
Phone clicks, contact form submissions and thank-you page views were instrumented as distinct events, so that a call and a form fill can be counted separately rather than collapsed into a single meaningless total. For a firm whose intake is mostly by phone, an untracked phone click is the whole business going unrecorded.
Restored the connection between channel and outcome
Once events fire correctly, conversions can be attributed to the channel that produced them. That is what turns analytics from a traffic report into a budget document: it becomes possible to say which source produced the calls rather than which source produced the most visits.
Watched it climb rather than declaring victory
Two conversions in the first month back is not a result, it is a signal that the plumbing works. Fifteen the following month and forty-nine the month after is the pattern you want, because it shows events firing consistently as more of the site gets instrumented.
What did not improve
The jump from 0 to 49 a month is not a 49-fold improvement in the firm’s marketing, and presenting it that way would be dishonest. Some of those conversions were always happening; they simply were not being counted. What changed is that the firm can now see them.
What the numbers genuinely show is a firm that spent five months making decisions blind and no longer does. The growth from 2 to 15 to 49 across March, April and May reflects both real improvement and a widening share of the site being measured correctly, and separating those two precisely would require data that does not exist for the dark period.
Questions about this case study
How can you tell broken tracking from a genuine drop in conversions?
By the shape. Real demand fluctuates: some weeks are better than others, and a bad month still produces something. Broken tracking produces exact zeros for long stretches while sessions carry on unchanged. Five months of perfect zeros against a thousand sessions a month is not a market condition.
Why does a phone click need to be tracked separately?
For service businesses where most enquiries arrive by phone, an untracked phone click means the main conversion path is invisible. If a channel drives calls but not form fills, and only forms are counted, that channel looks worthless and tends to get cut.
How often should conversion tracking be checked?
Monthly, and the check should be for zeros and flat lines rather than for the total. Most broken tracking is discovered months late, by which point the reporting has already been used to make decisions.
If this sounds like your situation
Fifteen minutes, no deck. I will look at your search data before the call and tell you which of these problems you have, and whether it is worth fixing. If it is not, I will say so.
Questions, answered.
What was Personal injury firm, Arizona starting point?+
Established personal injury practice in Phoenix and Mesa with an existing site. We took an organic foundation at 301 monthly clicks and lifted it to 358 in 90 days, plus expanded total impressions from 158k to 207k.
How does NetPageTwo work for law firms specifically?+
Practice-area page rebuilds in answer-first format, schema deployment (Organization, Person, LegalService), state-specific long-tail content, phone CTA optimization, and AI search citation work for state-level personal injury queries.
Did this work for both bilingual and English markets?+
Personal injury firm, Arizona was English-primary. Bilingual personal injury firm, run by the same firm leadership, was the bilingual case. Both case studies show different shapes of work for different markets.