The spin-off outdrew its parent and never lifted it.
Robert and Michelle King built The Good Wife, The Good Fight and Elsbeth from one set of characters. We tested whether launching the new show sent people back to the old one, and added Suits and Suits LA as a second family. In 0 of 6 premiere months did the parent clear a 20% lift bar.
No premiere lifted its parent. The best month was +14%; the worst was -42%.
The spin-offs drew far more attention than the originals. Elsbeth peaked at 550,000 searches a month (Apr 2024) against The Good Wife’s best of 201,000. Suits LA reached 1,220,000 in Feb 2025 while Suits sat near 40,000.
The decision this is for
Whether to fund a spin-off, or any extension of a hit, partly on the promise that it will revive the original. The buyer is a head of content or a brand lead with a successful product and a budget for the next one.
If the halo is real, the new product's case includes revenue on the old one. If it isn't, the new product has to pay for itself. Most extension cases count the halo; this is a test of whether it shows up.
The premiere months
| Parent | Premiere | Month | Searches | Prior-3 mean | Change |
|---|---|---|---|---|---|
| The Good Wife | Elsbeth season 1 | Feb 2024 | 90,500 | 97,000 | -7% |
| The Good Wife | Elsbeth season 1, month 2 | Mar 2024 | 110,000 | 97,000 | +13% |
| The Good Wife | Elsbeth season 2 | Oct 2024 | 110,000 | 118,333 | -7% |
| The Good Wife | Elsbeth season 3 | Oct 2025 | 135,000 | 118,333 | +14% |
| Suits | Suits LA | Feb 2025 | 40,500 | 44,033 | -8% |
| Suits | Suits LA, month 2 | Mar 2025 | 27,100 | 46,500 | -42% |
The Good Wife did grow: from about 73,000 searches a month in 2019 to about 141,000 in 2025. It grew slowly and never in a premiere month. That could be a slow halo, or a streaming deal putting the old show in front of new people; this test cannot tell them apart. What it rules out is the premiere-month halo most launch plans assume.
By Robert King’s account, CBS’s own focus groups found few people knew The Good Wife or The Good Fight, so Elsbeth was built to need neither.
Why people watch each show
| Reason | The Good Fight | Elsbeth |
|---|---|---|
| Politics | 8 | 0 |
| Writing | 6 | 1 |
| Representation | 5 | 0 |
| Characters | 2 | 10 |
| Format | 1 | 5 |
The Good Fight is watched for what it says; Elsbeth for who she is. The two lists barely overlap. The edgier show went to a paid streaming channel and the warmer one to free broadcast. What carried between them was a character, not a tone.
What AI assistants say about the family
| What the assistant did | Times | Stable? |
|---|---|---|
| Leaves Elsbeth off a list of the Kings' shows | 3 of 3 | Same every run |
| Recommends Elsbeth to fans of The Good Wife or The Good Fight | 0 of 4 | Same every run |
| Links The Good Wife and The Good Fight | 4 of 4 | Same every run |
| Sends Elsbeth fans back to the older shows | 1 of 2 | Changed between runs |
| Names representation as a reason to watch The Good Fight | 1 of 2 | Changed between runs |
| Cites CBS or Paramount+ | 1 of 14 | Same every run |
The links run backward. Asked what the Kings have made, the assistant left out the show now in its fourth season every time. It connects the old shows to each other, but never recommends the new one to their fans. And the owner’s own sites were cited once in 14 answers.
What this cannot tell you
Search is curiosity, not viewing. A lift in viewing with no lift in searches is not ruled out.
Google rounds these figures in steps of about 22%, so the 20% bar sits close to one step. A lift smaller than one step can read as zero. The direction across six months is the finding, not any single month.
Two families, one network’s shows. Reviews are critics, not the median viewer. The assistant is one model, asked two or three times.
How it was built
Three questions, each with its predictions written down first. Search: parent US monthly volume, 2019 to September 2026, in each premiere month against the three months before. Reasons: 25 professional reviews and creator interviews, each coded to five reasons registered in advance. Assistants: six questions to ChatGPT with web search on, asked twice, with a third run of the portfolio question. Only patterns that repeated are stated as findings.
| Prediction | Registered | Result |
|---|---|---|
| Arm 1 · a premiere lifts the parent (halo) | 45% | Miss |
| Arm 1 · the parent is independent of premieres | 45% | Hold |
| Arm 2 · six registered reasons-to-watch priors | — | 4 hold, 2 miss |
| Arm 3 · assistants connect the family | 65% | Hold |
Nothing was commissioned and no studio was contacted. Runs d151744d · 9b6f4d3d · db6ee264 · ece8668b.
Read the story version
The Sound Decisions post draws the brand lessons from the same data.
Whether to fund the next show is one of six decisions
This study tests one assumption in a content budget. The Streaming Decision Brief takes the six decisions a streaming executive owns and answers each from filings, with the limits of the evidence stated.
See the six decisions$495 standard, $995 custom. Pre-made and delivered on payment. If the decision turns out not to be decidable from evidence, it is refunded.