SerpApi vs DataForSEO: an independent benchmark on 30 Italian keywords
On 30 keywords searched from Italy, SerpApi finds 59% of the organic results seen in the browser, DataForSEO 33%. But neither is stable: repeat the same search and about half the page changes.
Davide Ruggeri
The result in brief
On 30 keywords searched from Italy, SerpApi finds 59% of the organic results I saw in the browser, DataForSEO 33%. SerpApi also respects the order better and gets the first result right more often, but in my test it cost about 14 times as much per search.
One figure matters more than the ranking: neither of the two is stable. Repeat the same search under the same conditions and about half of the page changes.
I needed the test to choose the SERP provider for a tool that checks keyword intent before suggesting keywords for Google Ads. I published the numbers because the comparisons you find around are almost all written by one of the two providers or by a competitor.
How I ran the test
On 14 October 2026 I searched 30 keywords on Google from Chrome, in an incognito window, from Calcinate in the province of Bergamo. A bookmarklet saved the results page as I saw it. Right after each of my searches, a script ran the same query with SerpApi and with DataForSEO.
The API settings were the same for both: Google Italy, Italian language, desktop, and the coordinates of my town as the location. To measure stability, each API repeated every search three more times, under the same conditions.
The keywords were split into six types, five per type (all searched in Italian):
informational: come fare il pane in casa, come togliere il calcare dalla doccia, quanto dura il passaporto, cos’è l’inflazione, come piantare i pomodori
commercial: miglior robot aspirapolvere, migliori cuffie wireless, migliore friggitrice ad aria, miglior assicurazione auto, miglior materasso memory
transactional: scarpe running uomo, noleggio auto aeroporto orio al serio, volo milano londra, bici elettrica offerta, prestito personale online
local: pizzeria, idraulico, farmacia di turno, ristorante sushi, meccanico vicino a me
navigational: agenzia delle entrate, trenitalia, orio center, inps, amazon
current events: meteo bergamo, atalanta, prezzo benzina oggi, sciopero treni, classifica serie a
The comparison covers the top 10 organic results. Ads, maps, videos and the like I treated separately as special elements. I have no commercial relationship with either SerpApi or DataForSEO.
How closely they match my SERP
SerpApi was closer to what I saw on almost every measure. It had the same first result as mine on 18 keywords out of 30, DataForSEO on 4. A result in exactly the same position as in my page happened in 26% of cases with SerpApi and 6% with DataForSEO.
The order of the shared results also holds up better with SerpApi. On a scale from −1 (reversed order) to 1 (identical order), SerpApi reaches 0.68 and DataForSEO 0.21.
[embedded content: Benchmark of 14 October 2026 · 30 keywords, 5 per type · top 10 organic results, first request of each API]
The clearest difference is on transactional queries, the ones that matter most to people running campaigns. DataForSEO is ahead only on current-events searches, where the page keeps changing anyway.
A curious detail: the two APIs also resembled each other very little, with only 25% of results in common. If they reproduced the same page well, they should look much more alike.
Stability: the same search, repeated
I repeated every search four times per API, under the same conditions. Between any two requests, SerpApi returned on average 49% of the same results, DataForSEO 40%. The first result stayed the same in 50% of cases with SerpApi and 42% with DataForSEO.
This puts the numbers of the previous section in perspective. On a single keyword, the share of my results found swung on average by 46 points between the best and the worst request with SerpApi, and by 60 with DataForSEO. On “miglior assicurazione auto”, for example, both went from 2 to 8 of my 9 results depending on the request.
Averaging over all requests, the gap narrows but remains: 58% for SerpApi, 45% for DataForSEO. DataForSEO’s first request had been unluckier than average.
I can’t say how much of this variability comes from Google and how much from the providers. In practice it makes little difference: a single data point from a SERP API should be treated as a sample, not a measurement.
AI Overview and special elements
Here DataForSEO regains ground. The AI Overview appeared in my page on 16 keywords: DataForSEO reported it on 11, SerpApi on 9. The cited sources matched mine in 70% of cases with DataForSEO and 75% with SerpApi.
On the other special elements (local pack, videos, images, products, “People also ask”) DataForSEO matched my page in 73% of cases, SerpApi in 60%.
If you care about the shape of the SERP more than who is in it, the gap between the two is much smaller than on organic results.
Costs and speed
In the test, a search cost about 0.028 dollars with SerpApi and 0.002 with DataForSEO in Live mode. On a thousand searches a month that comes to about 28 dollars against 2.
The comparison isn’t entirely like for like. SerpApi sells monthly plans, and the cost per search drops with larger plans. DataForSEO works on top-ups, with a minimum deposit of 50 dollars, and costs even less if you accept receiving results in a queue after a few minutes (pricing details).
On speed I found no consistent difference. In the final run the average times were similar, 6.4 seconds for SerpApi and 6.0 for DataForSEO. In an earlier run SerpApi took 11.1 seconds on average and DataForSEO 5.1.
Limits of the test
It’s a small test and should be read as such.
Thirty keywords, one day, one place. With other queries or from another city the numbers would change.
My search is a single sample, while the APIs were queried four times. Given how much the SERP varies, my page could also have been just one of many possible ones.
Google also locates me from my IP, the APIs only from the coordinates I pass. On local queries part of the difference comes from here.
I don’t compare ads. The bookmarklet that recorded my page stopped recognising them after Google changed its HTML structure, and I’d rather not publish a figure I’m not sure of.
I match my results to the APIs’ by site and title, because Google in the browser no longer shows the direct address of links. Some matches may be wrong.
Which one I chose and why
For my tool I chose SerpApi for SERPs. It has to check keywords that will end up in Google Ads campaigns, so mostly transactional and commercial ones, and that’s where SerpApi was most faithful. The cost per search is higher, but for the volume I need it remains sustainable.
DataForSEO stays in the stack, though, for search volumes, CPC and keyword ideas. For that data there’s no need to imitate the page a person sees, and price makes the difference.
If your case is different, the choice may change. To monitor thousands of keywords every day, cost matters a lot, and DataForSEO becomes hard to beat. If you’re mostly interested in special elements or the AI Overview, in my test DataForSEO was also a bit more accurate.
If you want to understand how to set up keyword research or campaigns for your site, tell us about your project.