Your company changed its story. The review sites, the directories, Wikipedia, and the AI assistants your buyers now ask first are still telling the old one. I measured how far behind they run.
I audited 47 B2B SaaS companies that renamed, rebranded, or publicly changed category between January 2024 and October 2025. Every one is publicly traded, PE-owned, or a late-stage scale-up; the list includes Docusign, Zoom, F5, UiPath, Box, Dayforce, and Corpay. As of July 2026, a median of 21 months after their announcements, 41 of the 47 (87%) still had at least one major public surface telling the old story. When I asked a frontier AI model what each company does, one in seven got described purely as the thing it had spent millions to stop being, and another 29 heard the old category first, with the new story trailing.
Five findings, each one self-contained:
- 41 of 47 repositioned B2B SaaS companies (87%) still had at least one major public surface telling their old story, a median of 21 months after the announcement (SandsDX audit, July 2026).
- One in seven of the 47 was described by an AI model purely as its old business, with web access off (SandsDX audit, July 2026).
- Companies had updated their own LinkedIn description in 89% of cases, while only 29% of their Wikipedia leads carried the new story (SandsDX audit, July 2026).
- 40% of the cohort’s G2 profiles did not lead with the new positioning (SandsDX audit of 47 B2B SaaS companies, July 2026).
- All 6 companies with a fully synced public story had changed their name; 0 of the 30 keep-the-name repositionings were fully synced (SandsDX audit, July 2026).
That gap has a name.
What repositioning lag is
Repositioning lag is the gap between announcing a new positioning and the moment the revenue system carries it. The company changes its story while the systems that deliver that story to buyers, across message, targeting, motion, and measurement, keep telling the old one. Pipeline tracks what the systems say, and the systems are behind.
The announcement is an event. The repositioning is a migration. Most companies run the event, skip the migration, and then spend two quarters arguing about whether the new positioning “worked.” The positioning never got deployed. The press release went out; the system kept shipping the old story on every surface a buyer touches.
Four systems, one lag
A positioning change has to propagate through four systems before pipeline reflects it:
- Message. Every surface that describes you: your site, review profiles, directories, Wikipedia, sales decks, outbound sequences, and now the AI models that answer for you when you are not in the room.
- Targeting. The accounts and personas the team actually pursues. New category, same old list, and the new story gets pitched to buyers selected for the old one.
- Motion. How you sell: qualification, demo flow, objection handling, pricing conversation. A team that demos the old product story inside a new category confuses exactly the buyers the repositioning was meant to win.
- Measurement. What you count. If the funnel stages, attribution, and forecast categories still describe the old business, every dashboard quietly pulls the team back to it.
The lag lives wherever the announcement never landed. The audit below measures the message system, because it is the one system you can score from the outside, and because it is the one AI assistants read.
The audit: 47 companies, four surfaces
The cohort: 47 B2B SaaS companies that renamed, rebranded, or changed category positioning between January 2024 and October 2025, spanning security, HR, payments, martech, adtech, developer tools, data, and vertical SaaS. Every company clears a scale bar: publicly traded, PE-owned, or private with roughly $100M+ raised or a $1B+ valuation. Seventeen changed their name entirely. Examples of the moves: Abnormal Security became Abnormal AI, Ceridian became Dayforce, MessageBird became Bird and moved from messaging APIs to CRM, 1Password declared a new category called Extended Access Management, Docusign repositioned from e-signature to Intelligent Agreement Management, Klaviyo moved from email marketing to B2C CRM, and Gong declared a Revenue AI Operating System category.
Each company got scored on four surfaces:
- A frontier AI model’s answer to “What does this company do?” with web access disabled, probed blind before any audit context existed, so nothing led the witness. This is the layer that answers when a buyer asks an assistant without search.
- LinkedIn company description.
- G2 profile description.
- Wikipedia article lead, where one exists.
Every surface got one of three verdicts: new story (leads with the new positioning), mixed (contains both, leads with the old category), or old story. Full methodology and limitations at the end.
Finding 1: you updated LinkedIn, and almost nothing else followed
| Surface | Leads with new story | Mixed | Old story only |
|---|---|---|---|
| LinkedIn (n=47) | 89% | 4% | 6% |
| G2 or equivalent review site (n=47) | 60% | 19% | 21% |
| Wikipedia (n=34 articles exist) | 29% | 24% | 47% |
| AI model answer (n=47) | 23% | 62% | 15% |
LinkedIn syncs because you own the edit button. The surfaces you do not own move at their own speed, and those are the surfaces AI models weight: review sites, Wikipedia, third-party coverage.
The concrete versions are worth staring at. Bird renamed from MessageBird in February 2024 and repositioned to CRM; in July 2026 its G2 description still opened with “MessageBird is a cloud communications platform.” TigerData’s Wikipedia article is still titled “TimescaleDB,” and Teads’ English Wikipedia article is still titled “Outbrain.” Zoom renamed itself Zoom Communications to mark its AI-first platform story; Wikipedia’s lead still defines it by videoconferencing. Two and a half years after 1Password launched Extended Access Management, its LinkedIn, G2, and Wikipedia descriptions all still said password manager. These are well-run companies with real marketing teams. The lag is structural, and it comes for everyone who treats the announcement as the finish line.
Finding 2: ask an AI what you do, and it answers from the old story
Seven of 47 companies got an AI answer that told only the old story. The one that made me sit up was Gong. Nine months after it declared the Revenue AI Operating System category, the model, which had no way of knowing this was a repositioning study, described a revenue intelligence platform that records and analyzes sales calls. Another 29 of 47 got a mixed answer: the old category first, the new story as a trailing clause. Only 11 of 47 led with the new positioning.
A word in defense of the mixed answers: a company mid-transition should be described with both stories, and an answer that carries both is fair. It is still expensive. The lead frame is the one buyers remember and repeat, and in 36 of 47 cases that frame was the old category. The model frequently knew a rename had happened and described the old business anyway. It had read the press release. It did not believe it.
The scale of this matters because the buyer journey now starts inside AI answers. In Gartner’s 2025 buyer research, B2B buyers complete most of their evaluation before ever contacting sales, and AI search has become a primary research surface. My GEO/SEO executive guide covers the shift in detail.
The loop that kills repositionings
The lag costs more than tidiness, and here is the mechanism. Unsynced surfaces keep generating demand shaped like the old company. Those leads book calls expecting the old product, and sales works them with the new pitch. Win rates dip, at exactly the moment scrutiny of the repositioning peaks. Two quarters in, the CFO looks at falling conversion and concludes the new positioning failed. What actually happened: the old story kept manufacturing old-company demand, and the new pitch lost it. The lag fabricates the evidence that gets used to kill the repositioning. Sound stories get rolled back on exactly this data, and nobody in the room checked which company the pipeline was still buying from.
The loop survives because propagation has no owner. Brand marks the launch done at day 30. Product marketing rotates to the next launch. RevOps was never told the ICP changed, and demand gen’s campaign taxonomy still carries the old category names. The launch retro happens at day 30. The lag shows up at day 300. No meeting exists where those two facts sit in the same room, so the miss gets filed under strategy instead of propagation.
Lag or disagreement
Some surfaces stay old because they are slow. Others stay old because they are unconvinced. Wikipedia’s editors follow neutrality rules that resist vendor category language by design. G2’s taxonomy tracks what buyers actually shop for, and buyers shop in categories rather than announcements. When those surfaces keep calling an agentic automation platform an RPA vendor, part of what you are seeing is the market declining to ratify a category the company declared.
The tell is your own customers. Pull your last ten win-loss notes and listen to the words. If customers describe you in the new language while third-party surfaces lag behind, you have a propagation problem, and the playbook below fixes it. If your own customers still describe you the old way, the story itself has not landed, and no amount of surface syncing will ratify a positioning the market does not believe. That is a different project, and it sits upstream of this one.
Finding 3: the rename paradox
Of the 17 companies that changed their name, 10 still had the old name prominent on a major surface: LinkedIn URLs still carrying the former name, Wikipedia articles still titled with it, review profiles opening with it. A rename without a surface-by-surface migration plan means buyers, search engines, and AI models keep resolving to the old entity.
Now the part I did not expect. Only 6 of 47 companies had every found surface plus the AI answer leading with the new story, and all 6 had changed their name: Dayforce, Corpay, Black Duck, Candescent, Windsurf, and Personify Health. Zero of the 30 companies that kept their name while changing category came out fully synced.
My read: a rename works as a forcing function. Changing the name creates a migration project with an owner, a deadline, and a surface inventory, and nobody downstream can treat it as optional. There is a rival explanation the data cannot rule out: companies willing to change their name may simply be running bigger, better resourced transitions, and carve-outs and mergers have no choice but to re-explain themselves everywhere. Both readings point the same direction for operators. The companies that got clean ran the story change as an operations project. The keep-the-name repositionings ran launches, and the old story survived wherever nobody was assigned to remove it. Three companies even had their own LinkedIn description still telling the old story.
How to measure repositioning lag: the 30-minute diagnostic
You can measure your own lag this afternoon. Score each item new / mixed / old:
- Ask two AI assistants “What does {your company} do?” with browsing off. Compare to your current homepage.
- Read your G2 (or category review site) description out loud next to your homepage hero.
- Read your Wikipedia lead, if you have one, and your Crunchbase description.
- Pull your three most-used outbound sequences. Which story do they pitch?
- Open the deck your sales team presented last week, as opposed to the one enablement published.
- Check your paid search ads and their landing pages.
- Look at the last 20 accounts sales worked. Do they fit the new ICP or the old one?
- Read your funnel stage definitions and forecast categories. Which business do they describe?
Eight items, four systems. Items 1 through 6 are message, item 7 is targeting, item 5 doubles as motion, item 8 is measurement. Count the non-new answers: that number is your repositioning lag, and each item is a work order. If most of your “old” answers cluster on surfaces you do not own, you have a visibility migration to run. If they cluster in sequences, decks, and stage definitions, the story never made it inside the building either.
For a number you can track quarter over quarter, invert it. The share of checked surfaces leading with the story you currently sell is your story sync rate. Across the four surfaces in this audit, the 47-company cohort averages about 52%. The six clean companies sit at 100%, and they are the ones whose category nobody argues about.
What closing the lag looks like
The fix is unglamorous and specific, which is why it works:
- Inventory every surface that describes you: owned, earned, and machine. There are more than you think. The audit above scored four per company; a real inventory for a mid-market B2B SaaS company runs to dozens of entries once you count directories, partner listings, review sites, app marketplaces, and the AI models themselves.
- Sync in the order buyers encounter surfaces rather than the order they are easy to edit. Review profiles, Wikipedia (through its editorial process, with citations), directories, and structured data on your own site typically move the AI answer more than another homepage revision.
- Migrate the motion: sequences, talk tracks, demo flow, and qualification against the new ICP. The audit measured public surfaces. In client work the inside surfaces are usually further behind.
- Move measurement last, and move it. Rename funnel stages, rebuild attribution around the new motion, and set the forecast to the business you are becoming. Teams sell what the dashboard rewards.
- Re-probe quarterly. Ask the assistants again. The answer drifts back toward the training data’s majority story unless the public record keeps corroborating the new one.
Appendix: the full table
Every verdict in the study. AI = the blind model answer. Scoring: NEW leads with the new positioning, MIXED contains both but leads old, OLD tells only the old story, N/A means the surface does not exist.
| Company (former name) | Change | Type | AI | G2 | Wikipedia | |
|---|---|---|---|---|---|---|
| Abnormal AI (Abnormal Security) | Apr 2025 | rename | MIXED | NEW | NEW | N/A |
| Black Duck Software (Synopsys SIG) | Oct 2024 | carve-out rename | NEW | NEW | NEW | NEW |
| TigerData (Timescale) | Jun 2025 | rename + category | MIXED | NEW | MIXED | MIXED |
| LevelBlue (AT&T Cybersecurity) | May 2024 | spin-out rename | NEW | NEW | MIXED | NEW |
| Absolute Security (Absolute Software) | Apr 2024 | rename | MIXED | NEW | NEW | NEW |
| Bird (MessageBird) | Feb 2024 | rename + category | OLD | NEW | OLD | N/A |
| Superhuman (Grammarly) | Oct 2025 | rename | MIXED | MIXED | NEW | MIXED |
| Personify Health (Virgin Pulse + HealthComp) | Feb 2024 | merger identity | NEW | NEW | NEW | N/A |
| 1Password | May 2024 | category | MIXED | OLD | OLD | OLD |
| Dashlane | Apr 2025 | category | OLD | NEW | OLD | OLD |
| Snyk | May 2025 | category | OLD | NEW | OLD | OLD |
| KnowBe4 | Nov 2024 | category | MIXED | NEW | OLD | N/A |
| Mimecast | Jul 2024 | category | MIXED | OLD | OLD | OLD |
| NetSPI | May 2024 | category | NEW | NEW | MIXED | N/A |
| Docusign | Apr 2024 | category | MIXED | NEW | MIXED | NEW |
| ZoomInfo | May 2025 | reposition | NEW | NEW | MIXED | OLD |
| Klaviyo | Feb 2025 | category | MIXED | NEW | NEW | NEW |
| Miro | Oct 2024 | reposition | OLD | NEW | NEW | OLD |
| Webflow | Oct 2024 | category | MIXED | NEW | NEW | OLD |
| UiPath | Apr 2025 | reposition | MIXED | NEW | NEW | NEW |
| Airtable | Jul 2025 | reposition | MIXED | NEW | NEW | MIXED |
| Box | Nov 2024 | reposition | MIXED | NEW | NEW | OLD |
| Exabeam (merged with LogRhythm) | Jul 2024 | merger, name kept | MIXED | NEW | NEW | OLD |
| Contrast Security | Aug 2024 | category | MIXED | NEW | NEW | N/A |
| SecurityScorecard | Oct 2024 | category | OLD | MIXED | MIXED | N/A |
| EDB (EnterpriseDB) | May 2024 | rename + category | MIXED | NEW | NEW | OLD |
| F5 | Feb 2025 | category | MIXED | NEW | MIXED | MIXED |
| Zoom (Zoom Video Communications) | Nov 2024 | rename + category | MIXED | NEW | NEW | OLD |
| Elastic | May 2024 | category | MIXED | NEW | NEW | OLD |
| Glean | Sep 2024 | category | MIXED | NEW | NEW | OLD |
| OneAdvanced (Advanced) | Apr 2024 | rename | NEW | NEW | NEW | MIXED |
| Automation Anywhere | Jun 2024 | category | MIXED | NEW | NEW | OLD |
| ThoughtSpot | Sep 2025 | category | MIXED | OLD | NEW | OLD |
| Dayforce (Ceridian) | Feb 2024 | rename | NEW | NEW | NEW | NEW |
| Corpay (FLEETCOR) | Mar 2024 | rename | NEW | NEW | NEW | NEW |
| Candescent (NCR Voyix Digital Banking) | Sep 2024 | carve-out rename | NEW | NEW | NEW | N/A |
| North (North American Bancard) | Aug 2024 | rename | MIXED | NEW | MIXED | NEW |
| Later (merged with Mavrck) | Jan 2024 | merger, name kept | MIXED | NEW | MIXED | N/A |
| Teads (Outbrain) | Jun 2025 | merger rename | MIXED | NEW | NEW | OLD |
| Totango (merged with Catalyst) | Jan 2025 | merger, name kept | OLD | NEW | OLD | N/A |
| Uniphore | Jun 2025 | category | MIXED | NEW | OLD | NEW |
| Gong | Oct 2025 | category | OLD | NEW | NEW | N/A |
| Qualified | 2025 | category | MIXED | NEW | NEW | MIXED |
| SmartRecruiters | Oct 2024 | category | MIXED | NEW | OLD | N/A |
| Cotality (CoreLogic) | Mar 2025 | rename | NEW | NEW | NEW | MIXED |
| Windsurf (Codeium) | Apr 2025 | rename + category | NEW | NEW | NEW | N/A |
| Intellistack (Formstack) | Jun 2025 | rename + category | MIXED | NEW | OLD | MIXED |
Methodology and limitations
Cohort selection came from public rebrand and repositioning announcements between January 2024 and October 2025, spanning security, HR, payments, martech, adtech, developer tools, data, and vertical SaaS, excluding any SandsDX client. Inclusion requires scale: a public listing, PE ownership, or roughly $100M+ raised or a $1B+ valuation, verified per company; three sub-scale companies from an earlier draft of the cohort were excluded under this bar before publication. The AI probe used one frontier model (Claude, knowledge cutoff January 2026) with web access disabled, asked blind in two batches, each before the model instance saw any cohort or scoring context. Surface descriptions were collected on July 21, 2026.
Verdicts are single-rater qualitative calls with evidence snippets retained for every row. The call I went back and forth on longest was SecurityScorecard: the blind answer described security ratings with supply chain risk monitoring, which brushes against the company’s Supply Chain Detection and Response story, and I scored it old because ratings led and response was absent. Reasonable people could score it mixed. That is the honest cost of single-rater scoring, and it is why every verdict in the appendix table above travels with a recorded evidence snippet.
This audit measures public-story sync; it does not measure pipeline outcomes, and no claim here should be read as “these companies’ revenue suffered.”
What I expect next
This page is wave one. I will re-run the probe and the surface audit quarterly and update the numbers here, with a correction noted if any verdict changes. My prediction for the October wave: the six clean companies stay clean, and fewer than five of the thirty keep-name repositionings join them. If the data says otherwise, you will read it here.
The story changed. The system that tells it to buyers is still running the old play. That gap is measurable, it has a name now, and at most companies it is nobody’s job, which is exactly why it persists.
If you repositioned in the last two years, run the eight-item diagnostic before your next board meeting. The number you get back is the distance between the company you announced and the company your buyers are still being sold.