If you’re reading this, you're mining review data.
Odds are, you’ve already covered the official APIs on the platforms you need data from, and either couldn’t find what you needed, or didn’t feel like building 10 clients for 10 different APIs.
If that’s so, welcome to a wonderful world of pain.
You’re in luck though. This short guide is meant to help you navigate through the different API options you have, and decide what option to go for based on your requirements and objectives.
I first dive a little into what different types of API are available, and how to decide what makes sense for you. I then go on to a list of 5 solutions that have been around for a while, and are known to be reliable with solid customer support.
The first two are what I call “high level” APIs, in the sense that they handle a lot of problems you’re going to face – such as deduplication, consistent review schema across platforms, language detection & translation, etc – while the last 3 are lower level, in the sense that they essentially dump scraping results. You’ll still get them as JSON, and won’t have to dig into proxy rotation and stuff like that, but you’ll need to get familiar with each platform’s schema and specificities.
Let’s dig in!
What makes a good review API?
Dealing with customer review data comes with a few specific challenges that review APIs will handle to different extents
Access and authentication
Let’s take an example. If you want to export Google reviews, you might want to use Google’s own APIs.
But gaining access to the My Business API requires a few things:
- First, you’ll need to submit an application with Google to be able to access those APIs. You’ll need to mention how many clients have requested it, etc. If you’re doing this as part of product development, it makes sense to go through the validation process. But if you just need programmatic access to your own Google reviews, it might not be the best route.
- Second, you’ll need your clients (or you) to authenticate their Google account. Which means you’ll need to anticipate how you’re going to store those credentials securely, have your clients grant access to their accounts with all the friction it implies, etc.
- Finally, you’ll need to scope the resources you have access to, and decide (and document for your clients) what to store and what to ignore. That last step might sound trivial, but you’ll likely run into pretty messy accounts, and most clients won’t give you full access without asking any questions.
Some third-party APIs don’t require access to serve review data.
However, if you need to interact with those reviews (reply for example), or if you need access to reviews that aren’t publicly available (behind a login for example), you won’t be able to work around the platforms’ official APIs & authentication logic.
Data schema
Integrating with any external system requires getting familiar with the way they store the data.
Things like:
- Pagination – page index, offset, tokens, etc –
- Deduplication – How are updates handled? Deletions? Replies?
- Translation – What language is the content available in? What language codes are used? How is the content translated, if at all?
If you only need reviews from a single review platform, it makes sense to go through those steps, and get acquainted with the API you’re working with.
But if you’re looking at reviews as “generic” objects across several different review platforms, this will put significant pressure on your development and maintenance cycles.
Here again, some third party APIs can provide consistent schema across all review platforms, while others will be lower-level, taking platform-specific params for each request, and returning data in a platform-specific schema.
The lower the API’s level, the more maintenance and development overhead.
Pricing models
Most review APIs have some sort of credit system, and ultimately charge you based on the number of “pages scraped” or based on the number of reviews you obtained.
At Reviewflowz, we do things slightly differently, and charge by review profile. A Review profile is a public review page for a product, location, or brand. For example, your local hairdresser on Google My Business, Instagram on the App Store, MacDonald's UK on Trustpilot, or reviewflowz on G2.
Here’s why. The number of reviews you end up with is ultimately pretty far from the value you’re getting from those reviews. And ultimately, the volume of reviews a business or brand receives is almost exclusively a function of their target market. The lower the price point, the higher the volume.
You’d typically need review data to conduct competitive analysis, to monitor reviews for a list of brands or products, or to run an ad-hoc analysis on the voice of the customer. The number of reviews is most often a poor indicator of the value you’d get from any of those processes.
Regardless, it’s important to understand what metric review APIs will use to decide how much to charge you to evaluate the initial price point, and how that price point might evolve over time.
Platforms supported
Unless you need data from extremely niche platforms, this is a little bit of a non-issue, depending on who you’re talking to.
First, “High level” APIs support most platforms. Once you support 20 platforms, it’s minimal effort to add support to add an extra one. The real challenge is in supporting the first 10 or 20.
As for low level APIs, you can almost always find one that supports the platform you need reviews from, although probably not in this list.
If you need data from very niche platforms, know that we commit to adding new platforms within 2 weeks for new clients on Reviewflowz – provided that the reviews are publicly available.
Extra services
This might feel overkill when you first read it, but hear me out.
If you’re digging into review data across more than 3 platforms, you’re in for a few surprises.
First, you’ll need to understand how to identify a listing, and what are the perma-IDs you can use to access the reviews you want to access.
For example, Google My Business has 5 different IDs you can use to access review data:
- Place IDs
- CIDs
- FIDs
- Location IDs
- Account IDs
Understanding what each of those does, and how to transform a user intention (or Job To Be Done) into one of those IDs so you can access the reviews takes a while.
Some APIs make this a lot easier by offering search functionality – basically transforming a brand name (or keyword of any sort) into a review listing on a given platform.
Some APIs even claim they can identify all the listings associated with a brand name, although from what I’ve seen most of those have stopped with that claim.
This will likely not come as a surprise, but search is HARD.
The way we handle this at Reviewflowz, is by providing a list of candidates. We don’t pretend to be able to infer the exactly right listing(s) from a keyword search, and ask our clients to provide intelligence (human, or artificial) to sort through the search results and select the right one(s).
#1 Reviewflowz
If you’re considering third party review APIs, we’ve basically built Reviewflowz for you.
No access of any sort is required to get reviews from any brand, on any platform.
The input & output schema are consistent across all platforms, reviews are deduplicated, updates are marked as such, removed reviews are marked as such, and replies (if any) are included in the review data.
We handle language detection and serve every review with consistent ISO 639-1 language code format. By default, reviews are served in the original language in which they were published, but you can opt into automated (open-AI powered) translation if you’re working on NLP use-cases. We also lemmatize reviews and can provide pre-lemmatized review content for light-weight NLP applications (for example real-time reports).
We price based on review profiles. A Review profile is a public review page for a product, location, or brand. For example, your local hairdresser on Google My Business, Instagram on the App Store, MacDonald's UK on Trustpilot, or reviewflowz on G2.
Use our search endpoint to find the right profiles, create them, and access reviews within minutes.
Our API plans start from $300 / month. Unit prices vary based on the frequency of updates you require, and on the overall volume of profiles you need.
By default, when you create a review profile on Reviewflowz, we’ll monitor that profile regularly and automatically update it with any new review – with built-in deduplication etc.
You can create webhooks to receive new reviews only as POST requests, avoiding polling & deduplicating logic.
For more information on our API plans, check out our documentation and reach out so we can chat about your exact requirements and see if we’re a good fit.
#2 Datashake
Datashake is another high-level review API, offering consistent schema, built-in deduplication, and access to reviews from almost all review platforms.
They also provide search functionality, but only for locations & e-commerce products.
They have unfortunately removed public pricing since last year, but at that time the entry price was around $500 / month.
It’s definitely a robust API to be able to access review data with a consistent schema across many different platforms.
The main drawbacks in my mind are that they’re known not to be very flexible to add new platforms (or rather, to try to upsell you into weird custom development plans) – this from prospects I’ve spoken with – and last I heard, their pricing was based on the # of reviews output, which makes it very hard to reliably predict your budget.
To learn more about Reviewshake, check out their documentation.
#3 Data For SEO
Now we’re starting to get into lower-level APIs.
DataForSEO basically provides a scraping output, but they do it reliably, and they do it cheap.
The main drawback is it can be extremely slow.
Their pricing is a little weird so I’ve just pasted a screenshot below for reference, but basically you can access up to 1M reviews for just $75.
You’ll have to handle language detection, deduplication, and figure out the whole place_id / cid mess, but if you’re only hoping to dump Google reviews one time, this might be a good direction.
They also support a few other review platforms such as Amazon, the App Store & Play Store, Trustpilot, and Tripadvisor.
Long story short, if you need lots of data just this once, and you need it cheap, this is definitely one to consider. The upside is they also have lots of SEO data, they do SERP scraping, backlinks, etc.
For more information, you can refer to their documentation.
#4 BrightLocal
BrightLocal are in the local SEO game.
They’ll be helpful for any kind of local business review. Think Google My Business, Yelp, Tripadvisor, and all other local directories on the internet.
They have a pretty cool APi feature that allows to search for a brand name across all directories and provide the (rare) ability to pass a webhook URL to get the results sent to you without having to poll for updates.
Of course, they also have the ability to fetch reviews across 80 local directories at $0.05 per request.
If you’re looking at review data for local businesses only, this is definitely one to consider. The main drawback is that they don’t currently support (and probably don’t plan on supporting) software review platforms, and 100% online review platforms like Trustpilot, the App Stores, etc.
You can learn more about BrightLocal on their API landing page here.
#5 BrightData
When in doubt, check out Brightdata.
Brightdata is probably as low level as it gets, but they do have tons of pre-built scraping engines and datasets that you can access, including lots of data around reviews.
Learn more about BrightData.
