Claude Haiku 5.5: Why Tallyroom trials don't convert

Scorecard

The analysis

Findings

1. Shortening the trial to 14 days (sales question) is not supported. Almost no lapsed teams say the trial was too long, and the time-related reasons point the other way: teams need more time to set Tallyroom up, not less.

   High confidenceBase 460

   Q2 'The trial was too long' = 0.2% (about 1 respondent). 'The trial was too short' = 1.5%. 'Forgot the trial was running' = 3.7%. Trial-length reasons combined = 5.4% (about 25 of 460). 'Didn't have time to set it up' = 18.7% overall and 27.4% of 1-9 employee teams. 24 of 109 open-text answers describe running out of time or setup taking too long, and several say the trial ended before the team had started. The 14-day option would shorten the window for the teams whose main barrier is time.

2. The largest fixable barrier is setup effort, not the integration gap or price. Assisted setup is the largest and cheapest pool to address, and it is concentrated in the smallest firms, which are 55% of the lapsed-trial population.

   Medium confidenceBase 460

   Unweighted, 'didn't have time to set it up' = 18.7%. Weighted to the population size mix (55/35/10), it is 21.6%, about 355 of 1,640 lapsed teams (0.55x27.4 + 0.35x15.9 + 0.10x10.0). Unweighted it is 27.4% for 1-9 employees, the single largest reason in that band. Open-text corroborates this (24 of 109 mention time or setup length, 14 of them in the 1-9 band). Caveat: Q5 does not test whether assisted setup would convert these teams, so the size of the conversion gain is unknown.

3. Sage integration (product question) has strong demand, but it is concentrated in Sage users and in the larger firms. It is not a general need. Sage-using teams are the most likely to say the integration is their blocker.

   High confidenceBase 117

   Q3: 23.7% of all respondents would need a Sage integration before paying (109 of 460; 95% margin of error +/-3.9 points, simple random sampling assumed). Among Sage users (n=117), 85.5% would need it, and 58.1% of Sage non-converters name no integration as their main reason (about 68 teams). Q3 need by size: 7.7% (1-9), 28.0% (10-49), 40.9% (50-249). Weighted to the population size mix, the need is 18.1%. Open-text names Sage in roughly 15 answers, including Sage 50, Sage 200, Sage Business Cloud and Sage Accounting.

4. The finance question (how many would pay if connected) can be answered only as stated intent, and it is the most optimistic number in the survey. Among teams blocked by a missing integration, 61.3% say they are likely to pay. Across the lapsed-trial population this is roughly 10% (about 164 of 1,640 teams). Real conversion would be lower.

   Low confidenceBase 93

   Q5 net likely (fairly or very likely to pay GBP 6 per user per month if connected) = 61.3% of the 93 'no integration' respondents (57 teams), against 37.6% of all 460. Weighted, 16.3% x 61.3% = about 10.0% of lapsed trials (about 164). Q5 is hypothetical, is asked at a fixed price, and is not broken out by software, so the 61.3% mixes Sage, FreeAgent and other users. Stated intent in purchase surveys usually overstates actual purchase.

5. Trial length is not the constraint and should be left at 30 days for now. The 'forgot' problem (3.7%) is cheap to fix with reminders and does not justify a shorter trial.

   High confidenceBase 460

   Trial-length reasons combined are 5.4%. 'Forgot' = 3.7% (about 17 teams), which a reminder sequence can address. Confidence is high that trial length is not the main lever. However, this survey covers only lapsed teams. It cannot show whether a shorter trial would lower or raise conversion among the teams that do pay, so this needs an experiment before any permanent change.

6. Price is a real but secondary reason and it does not differ significantly by company size. The complaints are mainly about per-user pricing, which makes per-company pricing the more relevant pricing test than a simply cheaper plan.

   Medium confidenceBase 460

   Q2 'Too expensive' = 15.9% overall: 19.0% (1-9), 13.2% (10-49), 15.5% (50-249). Two-proportion z-test for 1-9 vs 50-249: z = 0.77, p = 0.44, so the difference is not significant. Weighted to the population mix, 16.6%. Q3 'per-company pricing (not per user)' = 20.4%. About 23 open-text answers mention price, nearly all about per-seat or per-user cost.

7. Approval workflow is a real blocker for the larger firms and should be treated as a segment-specific issue. It is not a general trial-to-paid driver.

   Medium confidenceBase 110

   Q2 'Approval workflow didn't fit' = 12.6% overall (10.3% weighted), but 21.8% of 50-249 employee teams (n=110) and 7.1% of 1-9 employee teams. Q3 'custom approval chains' = 22.0%, and 27.3% of 50-249 employee teams. About 15 open-text answers describe approval routes that Tallyroom does not support, such as multiple approvers by expense type, outside approvers, or batch sign-off.

8. A meaningful share of lapsed teams had no real need for an expenses tool. No product change will convert these teams, so they set a ceiling on conversion.

   Medium confidenceBase 460

   Q2 'Decided we didn't need it' = 10.0% overall, 13.7% of 1-9 employee teams (about 12% weighted). Q3 'None of these' = 36.5% (40.5% of 1-9 employee teams). About 12 open-text answers say expenses are rare or they use a company card. Note that this is a stated, not a tested, lack of need.

9. Receipt capture quality and mobile upload reliability are a trust problem for some teams. This is a monitoring item rather than a lever for next quarter.

   Medium confidenceBase 460

   Q2 'Problems with the mobile app' = 6.3%. About 10 open-text answers describe OCR misreads, cropped receipts or failed uploads, including a misread total that led a team to stop trusting the app. Q3 'better receipt scanning' = 15.0%.

10. The survey sample is not representative of the lapsed-trial population in company size, so all weighted figures are approximations and must be read as such.

    Medium confidenceBase 460

    Sample size mix is 36.5% (1-9), 39.6% (10-49), 23.9% (50-249) against a population of 55%, 35%, 10%. The sample has no 250+ firms, although the population shows none either. The response rate is 28.0% (460 of 1,640), with a voucher incentive. Weighting assumes that respondents within each size band resemble non-respondents, which is unlikely.

Open-text themes it coded

Missing integration with accounting software (Sage, FreeAgent, KashFlow, Clear Books, others) 22Not enough time to set up, or setup too long or too involved 24Trial ended before the team got going 9Price or per-user pricing model 23Approval workflow does not fit how the business signs off 15Receipt scanning or mobile app reliability 10No real need for an expenses tool (low volume, company card, other process) 12Chose a competitor or another tool 8

The survey it planned

4 screening questions and 10 questions, as the model wrote them.

1. S1

   Are you the main admin for your company's Tallyroom account?

   One answer
   • Yes
   • No

   Continues if Yes

2. S2

   Is your company based in the UK?

   One answer
   • Yes
   • No

   Continues if Yes

3. S3

   Did your company's Tallyroom free trial end in the last 90 days without your company starting a paid plan?

   One answer
   • Yes, the trial ended and we have not paid
   • No, we are still on the trial
   • No, we started a paid plan
   • I'm not sure

   Continues if Yes, the trial ended and we have not paid

4. S4

   How many people does your company employ across all of its sites?

   One answer
   • 1 to 9
   • 10 to 49
   • 50 to 249
   • 250 or more
   • I'm not sure

   Continues if 1 to 9; 10 to 49; 50 to 249

5. Q1

   Which of these, if any, were reasons your company did not start a paid plan? Select all that apply.

   Any that apply
   • The price was too high for our team
   • We were not sure it was worth paying for yet
   • We did not have time to finish setting it up
   • It did not work with our accounting software
   • A feature we needed was missing
   • Our staff did not use it
   • We did not have enough expenses to need an app
   • We chose a different tool
   • The decision needed budget approval that we did not get
   • Other (please specify)
6. Q2

   Thinking about everything that happened, which one reason mattered most in your company's decision not to pay?

   One answer
   • The price was too high for our team
   • We were not sure it was worth paying for yet
   • We did not have time to finish setting it up
   • It did not work with our accounting software
   • A feature we needed was missing
   • Our staff did not use it
   • We did not have enough expenses to need an app
   • We chose a different tool
   • The decision needed budget approval that we did not get
   • No single reason; the trial just lapsed
7. Q3

   Which accounting software does your company use? Select all that apply.

   Any that apply
   • Xero
   • QuickBooks
   • Sage
   • FreeAgent
   • Excel or spreadsheets
   • Something else
   • We do not use accounting software
8. Q4

   Which of these, if any, would your company have needed Tallyroom to do before it would pay? Select all that apply.

   Any that apply
   • Connect to an accounting package we use that Tallyroom does not support yet
   • Match card or bank transactions to receipts
   • Handle VAT correctly on expense claims
   • Handle expenses in more than one currency
   • Handle mileage claims
   • Pay staff back for expenses through payroll
   • Set custom approval rules
   • Capture receipts when offline
   • Other (please specify)
   • None of these
9. Q5

   How important was it that Tallyroom connect to your accounting software?

   Scale
   • 1
   • 2
   • 3
   • 4
   • 5

   Scale 1-5: Not at all important to Essential before we would pay

   Judges: assumptive

   Assumes the company uses accounting software, although Q3 allows respondents to say it does not and no display logic excludes them from Q5.

   Assumes the company uses accounting software. Q3 offers 'We do not use accounting software', but no display logic routes those respondents past Q5.

10. Q6

    How long did your company take to decide whether to pay for Tallyroom?

    One answer
    • We knew within the first week
    • It took about one to two weeks
    • It took about three to four weeks
    • We never really decided; the trial just ended
    • The trial ended while we were still deciding
11. Q7

    On roughly which day of the 30-day trial did your company last approve or submit an expense in Tallyroom? Enter a number from 1 to 30. If you are not sure, enter your best guess.

    Number
    Judges: assumptive

    Assumes the company approved or submitted an expense during the trial. Companies that never did so must nevertheless enter a day from 1 to 30, producing a false activity date.

    Assumes the company approved or submitted at least one expense during the trial. Teams that never used it have no valid answer and are pushed into guessing a day.

12. Q8

    Looking back, how did the length of the free trial affect your company's decision to pay?

    One answer
    • It was too short to test Tallyroom properly
    • It was about the right length
    • It was too long, and we lost track of it
    • The length made no difference to our decision
    Judges: double-barrelled

    The option 'It was too long, and we lost track of it' combines two ideas: the trial being too long and the team losing track of it. A respondent who agrees with only one of them has no fitting answer, which muddies the test of the head of sales' view.

    The option 'It was too long, and we lost track of it' is double-barrelled. A respondent might think it was too long but not lose track, or vice versa.

13. Q9

    When you compared Tallyroom's price of £6 per user per month with what you expected, what did you think?

    One answer
    • Much higher than I expected
    • Somewhat higher than I expected
    • About what I expected
    • Lower than I expected
    • I did not think about the price
14. Q10

    Which of these would most likely have got your company to a paid plan?

    One answer
    • Help from the Tallyroom team to set up and connect the app
    • A cheaper plan for small teams
    • A shorter free trial
    • A longer free trial
    • Connection to an accounting package we use that Tallyroom does not support yet
    • Nothing would have changed our decision