Knotch Adaptive Content Experience for Zillow Rentals
V2 improved on 7 of 8 measures and slipped on 1. Visitors picked a persona about twice as often, used the chat about three times as often, bounced less, and converted more.
Visitors who picked a persona
ControlNo persona option
Visitors who used the chat
ControlNo chat
Sessions with one interaction; lower is better
Control44.0% to 25.8%
Visitors who converted
Control13.4% to 22.0%
Every measure for the ACE page, V1 against V2.
V2 led V1 on 7 of 8 measures. Control’s V1 to V2 change is shown on the right for context; conversion and bounce rate also improved on the Control page.
| V1 | V2 | Change | Control, V1 to V2 | |
|---|---|---|---|---|
| Using ACE Features | ||||
| Personalization Rate | 11.7% (188 of 1,604) | 24.8% (151 of 608) | +13.1 pts | No persona option |
| Chat Use | 4.5% (72 of 1,604) | 13.7% (83 of 608) | +9.2 pts | No chat |
| Engagement | ||||
| Time on Page | 4m 40s | 4m 39s | −1s | 1m 03s to 32.7s |
| Scroll Depth | 61% (1,604 visitors) | 64% (608 visitors) | +3 pts | 28% to 23% |
| Bounce Rate | 38.1% (898 of 2,358) | 19.8% (188 of 948) | −18.3 pts | 44.0% to 25.8% |
| Sessions per Visitor | 1.47 | 1.56 | +0.09 | 1.21 to 1.25 |
| Views per Visitor | 1.71 | 1.82 | +0.11 | 1.29 to 1.40 |
| Conversion | ||||
| Conversion Rate (Multi-Session) | 7.9% (127 of 1,604) | 17.8% (108 of 608) | +9.8 pts | 13.4% to 22.0% |
What changed in V2, and how visitors used it.
V2 launched on Sep 9 with four main changes. Two of them, the new hero and the new persona picker, put personalization front and center.
A typical answer now takes about 11 seconds, down from about 25.
Related MeasureChat Use
ACE builds each answer page from the components that best fit the question, instead of a fixed template.
Related MeasureChat Use, Time on Page
The top of the page puts more weight on personalized conversion paths.
Related MeasureConversion Rate (Multi-Session)
Picking a persona was redesigned for both mobile and desktop.
Related MeasurePersonalization Rate
24.8% of V2 visitors picked a persona, more than double V1’s 11.7%. The rise held on both desktop and mobile, so a shift in device mix doesn’t explain it.
Share of ACE visitors who picked any persona, compared with V1
Persona visitors as a share of ACE visitors on each device, compared with V1
| V1 | V2 | |
|---|---|---|
| Desktop | 25.7% (182 of 708) | 36.3% (147 of 405) |
| Mobile | 4.7% (42 of 896) | 16.3% (33 of 203) |
Desktop grew from 44% to 67% of ACE visitors. At V1’s device mix, V2’s personalization rate would still be about 25%, compared with 14% in V1.
In both versions, ACE visitors who used the chat converted more often than those who didn’t. Chat users also stayed much longer and rarely bounced.
Share of ACE visitors who used the chat, compared with V1
Chat users compared with non-chat visitors in the same version
| Conversion Rate (Multi-Session) | Time on Page | Bounce Rate | ||||
|---|---|---|---|---|---|---|
| V1 | V2 | V1 | V2 | V1 | V2 | |
| Chat Users small sample | 19.4% (14 of 72) | 31.3% (26 of 83) | 8m 57s | 8m 48s | 5.7% (23 of 402) | 7.4% (21 of 284) |
| Non-Chat Visitors | 7.4% (113 of 1,532) | 15.6% (82 of 525) | 3m 11s | 2m 39s | 44.7% (875 of 1,956) | 25.2% (167 of 664) |
| Control, for Context | 13.4% (274 of 2,038) | 22.0% (168 of 763) | 1m 03s | 32.7s | 44.0% (1,085 of 2,464) | 25.8% (246 of 952) |
ACE against Control, persona groups, and traffic.
ACE beat Control on all five engagement measures in both V1 and V2. ACE trailed Control on conversion in both versions.
| V1 ACE | V1 Control | V2 ACE | V2 Control | |
|---|---|---|---|---|
| Time on Page | 4m 40s | 1m 03s | 4m 39s | 32.7s |
| Scroll Depth | 61% (1,604 visitors) | 28% (2,038 visitors) | 64% (608 visitors) | 23% (763 visitors) |
| Bounce Rate | 38.1% (898 of 2,358) | 44.0% (1,085 of 2,464) | 19.8% (188 of 948) | 25.8% (246 of 952) |
| Sessions per Visitor | 1.47 | 1.21 | 1.56 | 1.25 |
| Views per Visitor | 1.71 | 1.29 | 1.82 | 1.40 |
| Conversion Rate (Multi-Session) | 7.9% (127 of 1,604) | 13.4% (274 of 2,038) | 17.8% (108 of 608) | 22.0% (168 of 763) |
ACE ÷ Control, V1 to V2
Control’s conversion rose too (13.4% to 22.0%), but ACE’s rose more, closing most of the gap.
Both pages bounced less in V2, and ACE’s edge grew. Lower is better.
The lead grew mostly because Control’s time fell (1m 03s to 32.7s).
The lead narrowed because Control rose.
In both versions, visitors who personalized converted at a higher rate than visitors to the Control page. In a single session, 11.2% of V1 persona pickers converted vs 10.5% on Control, and 25.8% vs 18.7% in V2; every persona also beat Control on multi-session conversion. Persona groups are small, so read the V1 to V2 changes as directional.
The Control page has no persona option, so every row is compared with all Control visitors.
| V1 | V2 | |
|---|---|---|
| Individual Landlord | 20.3% (30 of 148) | 27.5% (33 of 120) |
| Property Manager/Single Family | 13.7% (7 of 51) small sample | 42.1% (16 of 38) small sample |
| Enterprise/Multi-Family | 24.0% (6 of 25) small sample | 40.9% (9 of 22) small sample |
| No Persona Selected | 5.7% (91 of 1,597) | 10.4% (60 of 579) |
| All Control Visitors | 13.4% (274 of 2,038) | 22.0% (168 of 763) |
Visitors who converted in the same session they viewed the page single-session
| V1 | V2 | |
|---|---|---|
| Picked Any Persona | 11.2% (21 of 188) | 25.8% (39 of 151) |
| ACE Page, No Persona Selected | 4.5% (72 of 1,597) | 9.3% (54 of 579) |
| All Control Visitors | 10.5% (214 of 2,038) | 18.7% (143 of 763) |
Traffic was lower on both pages in V2, and desktop’s share of visitors rose on both, more on ACE.
Unique visitors ÷ days in each version
Desktop visitors ÷ all visitors
Definitions and notes on the data.
Every question asked in the ACE chat from Jun 30 to Oct 18, 2026. Visitors have asked 1,072 questions across 497 conversations, and the total grows every week.
Jun 30 – Oct 18, 2026
2,140 messages in total
Average across all personas
643 of 1,072 questions
Running total of questions, by week. Weekly detail begins Jul 27; the dashed line covers Jun 30 – Jul 26 (345 questions), which was reported as one block.
The themes and wording of visitors’ questions.
Most questions are about running a rental: leasing and operations, tenant screening, listing on Zillow, and marketing. Enterprise visitors ask mostly about marketing and advertising.
Share of each group’s questions, by the persona selected when the question was asked
Visitors ask practical, how-to questions about Zillow’s rental tools, from Rental Manager and screening to Rent Connect and 3D Home. Examples are the most frequent wordings in each theme.
Since Oct 5, visitors typed 73% of their questions in their own words; the rest came from prompt chips. Visitors with no persona selected leaned on prompt chips most.
Questions since Oct 5, when question type began being recorded (128 of 1,072 questions)
Questions visitors typed themselves, one per theme
Since Oct 5, all personas
Who asks, and how long their conversations run.
Individual Landlords and visitors with no persona selected ask 76% of all questions. A typical conversation runs about 4.3 messages, roughly 2 questions and their answers.
Questions asked, by the persona selected at the time
Messages include both visitor questions and ACE’s answers
| Conversations | Messages | Messages per Conversation | |
|---|---|---|---|
| Individual Landlord | 189 | 793 | 4.2 |
| Property Manager/Single Family | 61 | 267 | 4.4 |
| Enterprise/Multi-Family | 54 | 248 | 4.6 |
| No Persona Selected | 193 | 832 | 4.3 |
| All Personas | 497 | 2,140 | 4.3 |
Notes on the data.
All ACE visitors from Jun 30 to Oct 18, 2026, single-session conversion. Visitors who picked a persona converted at 18.4%, compared with 6.7% for those who didn’t, and above Control’s 12.7%.
57 of 310 visitors
124 of 1,852 visitors
Did vs didn’t personalize
Single-session: visitors who converted during the same visit in which they viewed the page ÷ unique visitors
Conversion for each persona, single-session and multi-session.
Visitors who personalized converted at a higher rate than visitors to the Control page. Across sessions, 27.1% of personalized visitors converted, compared with 15.7% on Control (18.4% vs 12.7% in a single session), and every persona beat Control on multi-session conversion. Half or more of Property Manager and Enterprise conversions came on a later visit.
| Conversion Rate (Single-Session) | Conversion Rate (Multi-Session) | |
|---|---|---|
| Did PersonalizePicked any persona, counted once | 18.4% (57 of 310) | 27.1% (84 of 310) |
| Individual Landlord | 17.8% (44 of 247) | 24.7% (61 of 247) |
| Property Manager/Single Family | 13.9% (11 of 79) small sample | 27.8% (22 of 79) small sample |
| Enterprise/Multi-Family | 7.7% (3 of 39) small sample | 30.8% (12 of 39) small sample |
| Didn’t PersonalizeACE visitors who never picked a persona | 6.7% (124 of 1,852) | 7.8% (144 of 1,852) |
| All ACE Visitors | 8.4% (181 of 2,162) | 10.5% (228 of 2,162) |
| All Control Visitors | 12.7% (353 of 2,776) | 15.7% (436 of 2,776) |
Visitors who used the chat converted at 26.4%, 2.8× the rate of ACE visitors who didn’t (9.4%) and above Control (15.7%). Chat users were 7% of ACE visitors but 45% of its Contact Sales converters.
Multi-session: converted in any visit during the period. Chat users and personas overlap, so these rows are separate views, not a split.
| All Conversions | Post Your Listing (Gated) | Post Your Listing (Ungated) | Contact Sales | |
|---|---|---|---|---|
| Chat UsersUsed the chat at least once | 26.4% (38 of 144) | 4.2% (6 of 144) | 13.9% (20 of 144) | 18.1% (26 of 144) |
| Non-Chat ACE VisitorsNever used the chat | 9.4% (190 of 2,018) | 1.9% (39 of 2,018) | 6.8% (137 of 2,018) | 1.6% (32 of 2,018) |
| Individual Landlord | 24.7% (61 of 247) | 5.3% (13 of 247) | 18.2% (45 of 247) | 6.1% (15 of 247) |
| Property Manager/Single Family | 27.8% (22 of 79) small sample | 5.1% (4 of 79) small sample | 6.3% (5 of 79) small sample | 25.3% (20 of 79) small sample |
| Enterprise/Multi-Family | 30.8% (12 of 39) small sample | 2.6% (1 of 39) small sample | 17.9% (7 of 39) small sample | 25.6% (10 of 39) small sample |
| All Control Visitors | 15.7% (436 of 2,776) | 9.0% (249 of 2,776) | 6.1% (168 of 2,776) | 1.5% (41 of 2,776) |
Share of chat conversations, by the persona selected at the time, Jun 30 – Oct 18
Conversion by persona for chat users is being added to reporting. Today’s view shows chat users as a group and which personas drive chat.
Gated listings, ungated listings, and Contact Sales.
Individual Landlords convert by listing; Property Managers and Enterprise visitors convert by contacting sales. Gated listing is the one conversion point where Control leads every persona.
Leading persona at each conversion point, Conversion Rate (Multi-Session), Jun 30 – Oct 18, 2026
Post Your Listing (Ungated)
Next: Enterprise/Multi-Family, 17.9%
Contact Sales
Next: Property Manager/Single Family, 25.3%
Post Your Listing (Gated)
Every persona trails Control here
Visitors who personalized beat Control on ungated listings and Contact Sales. Gated listings is the one type where Control leads every ACE group.
Single-session: converted during the same visit in which they viewed the page
| Post Your Listing (Gated) | Post Your Listing (Ungated) | Contact Sales | |
|---|---|---|---|
| Did PersonalizePicked any persona, counted once | 2.9% (9 of 310) | 11.9% (37 of 310) | 5.5% (17 of 310) |
| Individual Landlord | 2.8% (7 of 247) | 13.8% (34 of 247) | 2.4% (6 of 247) |
| Property Manager/Single Family | 2.5% (2 of 79) small sample | 2.5% (2 of 79) small sample | 11.4% (9 of 79) small sample |
| Enterprise/Multi-Family | 0.0% (0 of 39) small sample | 2.6% (1 of 39) small sample | 5.1% (2 of 39) small sample |
| Didn’t PersonalizeACE visitors who never picked a persona | 1.3% (25 of 1,852) | 4.6% (85 of 1,852) | 1.0% (19 of 1,852) |
| All ACE Visitors | 1.6% (34 of 2,162) | 5.6% (122 of 2,162) | 1.7% (36 of 2,162) |
| All Control Visitors | 7.4% (206 of 2,776) | 4.3% (120 of 2,776) | 1.3% (35 of 2,776) |
Over the full period, Property Manager and Enterprise visitors converted on Contact Sales at many times Control’s rate. Personalized visitors converted on ungated listings at 2.9× Control’s rate.
Multi-session: converted in any visit during the period, including the visit when they viewed the page
| Post Your Listing (Gated) | Post Your Listing (Ungated) | Contact Sales | |
|---|---|---|---|
| Did PersonalizePicked any persona, counted once | 5.2% (16 of 310) | 17.7% (55 of 310) | 11.6% (36 of 310) |
| Individual Landlord | 5.3% (13 of 247) | 18.2% (45 of 247) | 6.1% (15 of 247) |
| Property Manager/Single Family | 5.1% (4 of 79) small sample | 6.3% (5 of 79) small sample | 25.3% (20 of 79) small sample |
| Enterprise/Multi-Family | 2.6% (1 of 39) small sample | 17.9% (7 of 39) small sample | 25.6% (10 of 39) small sample |
| Didn’t PersonalizeACE visitors who never picked a persona | 1.6% (29 of 1,852) | 5.5% (102 of 1,852) | 1.2% (22 of 1,852) |
| All ACE Visitors | 2.1% (45 of 2,162) | 7.3% (157 of 2,162) | 2.7% (58 of 2,162) |
| All Control Visitors | 9.0% (249 of 2,776) | 6.1% (168 of 2,776) | 1.5% (41 of 2,776) |
Definitions and notes on the data.
ACE builds every answer from Zillow’s own articles. Three in four were posted by August 2020, and the average article is 7.8 years old. The most-cited article, landlord resources from the COVID-19 period, is cited because it gives a general overview of Zillow’s virtual tools; no newer general guide has been published since.
Since original post date
Three in four articles
One in ten articles
Share of the article library ACE draws on, by original post date
The Zillow articles ACE links to in its answers.
ACE’s answers most often cite Zillow’s landlord resources, listing guide, and customer stories. Citations are the times ACE linked an article in an answer, Jun 30 – Oct 18.
All personas
By the persona selected when ACE cited the article
“Close” marks an article within a few citations of the next one, so its place could shift with complete counts.
Notes on the data.