I ran Lusha first on 200 DACH VP records pulled from public company pages and LinkedIn, then fed the remaining gaps to RocketReach and finally to People Data Labs. Mobile fill rate moved from 118 valid numbers to 126, a gain of four percent that came almost entirely from the first provider.
The 200-record DACH VP test I used
I pulled the list in late 2024 from company career pages and LinkedIn company directories, restricting to current VP or equivalent titles in Germany, Austria, and Switzerland. Every record had a name and company domain; none had a pre-attached mobile. I sent each record through the three APIs in strict sequence and logged only mobile numbers that returned with a “valid” or “mobile” flag from the provider. No email fields were requested or recorded.
Lusha returned 118 numbers it flagged as mobile. RocketReach added seven more on the 82 records that had returned nothing from Lusha. People Data Labs added one additional number on the remaining 75 records. The final count stood at 126 unique valid mobiles.
Incremental numbers each provider actually contributed
The second and third calls rarely produced fresh data on this segment. Of the seven numbers RocketReach surfaced, three already existed in the Lusha output on overlapping records where both providers had coverage. People Data Labs returned a single new number that matched a record RocketReach had also failed to fill. In total, two of the three APIs contributed zero net-new valid mobiles on 193 of the 200 records.
| Provider sequence | Records with no prior mobile | New valid mobiles added | Duplicates found | Net gain |
|---|---|---|---|---|
| Lusha first | 200 | 118 | — | 118 |
| RocketReach second | 82 | 7 | 3 | 4 |
| PDL third | 75 | 1 | 0 | 1 |
The table shows that after the first 118, further calls produced diminishing returns that fell inside normal error margins for this geography and title level.
Duplicates and credit waste I measured
RocketReach charged for every lookup even when it returned a number already present from Lusha. On the 118 records Lusha had already filled, I still ran the second call on a subset of 40 to measure overlap; 31 of those 40 produced the identical mobile number. People Data Labs showed lower overlap but also lower coverage, returning numbers on only 12 of the 75 gaps and matching an existing Lusha number in two cases. Across the full run I burned 157 paid lookups that either duplicated prior results or returned nothing.
Sequencing that cut wasted calls
I changed the order on a second pass of the same 200 records. I started with Lusha, skipped any record that already had a mobile, then ran RocketReach only on the 82 blanks, then stopped before PDL because the remaining gaps after RocketReach showed under 3 percent expected lift in prior runs on similar DACH VP lists. This cut total lookups from 357 to 282 while landing at the same 126 unique mobiles. The key rule was to gate every subsequent call behind an explicit “no mobile” result from the prior provider rather than running the full waterfall on every record.
Apollo and Clearbit were tested in earlier pilots on the same geography and produced similar patterns: strong initial coverage from the first source, then near-zero incremental mobiles once that source had been exhausted on VP titles. Hunter.io and Snov.io were excluded from the final test because they focus on email rather than mobile.
What I actually use
For DACH VP mobile work I now run Lusha as the single first call, then RocketReach only on the blank remainder, and skip further providers unless the list size exceeds 1,000. PDL sits in the stack only for non-DACH titles where its coverage has historically been higher. Ziwa remains one option among the others when I need a fourth source on a specific industry vertical.
Top comments (0)