Published so you can check our work
How we score dialers
The rubric, rendered from the code that applies it. How pricing is verified and archived. What has and has not been tested. How the ranking moves under other weightings.
The rubric
Every dialer is scored out of 5 on five criteria. The table below is rendered from the same object the scoring function reads, so published and applied weights cannot differ.
| Criterion | Weight | What it measures |
|---|---|---|
| Value for money | 30% | Total cost of ownership at 1, 3, 10 and 25 seats, including telephony, against capability delivered |
| Dialing engine | 25% | Time between calls, line count, answering-machine detection, disposition speed, number rotation |
| CRM and data | 20% | Whether the tool can be a system of record, or needs one behind it |
| Setup and time to value | 15% | Hours from signup to a rep making calls, and how much of it needs an admin |
| Support and terms | 10% | Contract flexibility, seat minimums, pricing transparency, support channels and hours |
Value carries the largest weight deliberately. In a category where two tools that place one call at a time from a list can differ by a factor of twelve in price, cost is the axis that changes decisions most. Whether that weighting structurally favours the cheapest tool is a fair question, and the sensitivity table below answers it with numbers rather than argument.
The total is computed from the five sub-scores at render time and cannot be typed in. A build-time check fails if a stored total ever appears in the dataset. Sub-scores are editorial judgements; each carries a derived one-line justification on the review, and each vendor's changelog records when and why any of them moved.
What has been hands-on tested
None of the 21 dialers on this site has been hands-on tested by us. Every review says so above its score. Sub-scores for the dialing engine, setup and support come from vendor documentation, published pricing and terms, and the recurring themes in large public review bases, and the review says which.
When a tool is tested, the record will state who tested it, on what date, with how many dials, on which carrier, and what was measured, and that record will appear on the review. Until then, nothing on this site claims a test happened.
How pricing is verified
Vendor pricing pages first. Every price is read from the vendor's own pricing page. The URL, the retrieval date, who retrieved it, the method, and a Wayback Machine snapshot of the page as it stood that day are recorded on the review and in the verification log.
Third-party guides second, and labelled. Where a vendor gates pricing behind a demo, we cite a dated third-party pricing guide and the field says so. As of 2026-09-20 that applies to Orum (gated behind demo), Nooks (gated behind demo), Kixie (removed during 2026), Vulcan7 (gated behind demo), Readymode (removed during 2026), Koncert (published without units), Dialpad Sell (removed during 2026).
We do not guess. A figure that cannot be traced to a source is recorded as not published and the page says so in words.
Quarterly re-verification, by script and then by a person. A script in the repository fetches every pricing page, takes a fresh Wayback snapshot, looks for each recorded tier price in the page text, and opens a pull request. Prices it finds unchanged are re-dated; anything it cannot find, or any page that redirected, is flagged for a human and is never re-dated automatically. The log records every run. The first run was 2026-09-20; the next is due within 100 days of that, and a build fails if any vendor's verification is older than that.
How we use G2, Capterra and Trustpilot
As signal about direction, not as our score. A 4.8 across 1,540 reviews tells you something. A 4.9 across 12 does not.
We read the complaints, not the averages. Recurring negative themes in a large review base are the most useful public data about a product and are what appear in our cons lists.
Third-party ratings never enter our structured data. Where a G2 figure is cited it appears in the page text with the count and retrieval date attached.
What disqualifies a tool
- No published pricing and no credible third-party figure. We still review it and say that.
- Prices with no unit attached. Koncert's pricing page shows numbers without saying per what. That is recorded as its own transparency state.
- Claims we cannot verify. A connect-rate improvement claimed without a methodology is reported as a claim.
Where the numbers come from
Telephony arithmetic is computed in code from Twilio's published US rate of $0.014 a minute, read 2026-09-04 and archived. The model is published with its assumptions and worked examples. It accounts for per-call rounding up to the whole minute and for a picked-up voicemail billing a full minute.
Benchmarks each carry a named source, year, sample size where disclosed, retrieval date and archive link, all listed on /sources. Where a figure is our own estimate it is labelled as an estimate with no sample, and where we could reach only a secondary report of a study it is labelled secondary.
Corrections
Errors are corrected and logged on /corrections with what was wrong and what changed. Write to abhixchawla@gmail.com.
How the ranking moves under other weightings
The same sub-scores, re-weighted five ways. Top five under each. Published whatever it shows.
| Weighting | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|
| Published rubric (30/25/20/15/10) | DialSheet 4.5 | Close CRM 4.0 | Myphoner 3.9 | PhoneBurner 3.7 | Apollo.io 3.7 |
| Value de-emphasised (20/30/20/15/15) | DialSheet 4.4 | Close CRM 4.1 | PhoneBurner 3.9 | Myphoner 3.8 | Readymode 3.7 |
| Dialing-first (15/40/15/15/15) | DialSheet 4.4 | Close CRM 4.1 | PhoneBurner 4.0 | Orum 3.8 | Kixie 3.8 |
| Equal weights (20/20/20/20/20) | DialSheet 4.4 | Close CRM 4.1 | PhoneBurner 3.9 | Myphoner 3.8 | Apollo.io 3.7 |
| Value ignored (0/40/25/20/15) | DialSheet 4.3 | Close CRM 4.3 | PhoneBurner 4.2 | Orum 3.9 | Nooks 3.8 |
Full table with all 21 dialers on Where DialSheet loses.
Frequently asked questions
- How does MRR Nerds score dialers?
- Each dialer is scored out of 5 on five weighted criteria: value for money at 30 percent, dialing engine at 25 percent, crm and data at 20 percent, setup and time to value at 15 percent, support and terms at 10 percent. The total is computed from the sub-scores when the page renders and cannot be typed in.
- Has MRR Nerds hands-on tested the dialers it reviews?
- No. None of the 21 dialers on the site has been hands-on tested by us as of 2026-09-20, and every review says so above its score. Scores rest on vendor documentation, published pricing and terms, and public review corpora.
- How often is pricing re-verified?
- A script re-checks every vendor pricing page each quarter, takes a fresh archive snapshot, re-dates only prices it finds unchanged, and flags everything else for a person. The build fails if any verification is older than 100 days. Every check is in the verification log with its snapshot.
- Does the weighting favour the cheapest tool?
- Value for money carries 30 percent, so a free tool gains from it. The sensitivity table on this page re-ranks all dialers under four other weightings, including one that ignores value entirely, so a reader can see where the ranking holds and where it does not.
Who writes this
Abhi Chawla previously ran the SDR agencies LeanICP and Meetings Machine. He writes MRR Nerds' sales dialer reviews and pricing guides, focusing on calling workflows, software costs and carrier charges.
Editorial policy, verification log, the dataset, corrections.