It's interesting to see this valuation on search as a service, given the leader, Elastic, is in obvious valuation trouble. The market is unimpressed: the Elastic stock hasn't net moved higher since mid January (so most of the time they've been public).
Just take a look at the actual business performance of Elastic.
$271m in sales for the last fiscal year. Negative $101 million operating loss. Pretty bad, although not extremely unusual for SaaS companies in high growth mode. So there must be great growth going on, right? No.
They added a mere $9m in sales last quarter. $89m in sales with a $42m operating loss (whoa). They added an additional $10m in operating loss and gained a mere $9m in sales.
So if they can keep up that rate of growth, they might generously add $45-$50 million in sales this year. Maybe 16%-18% growth for a company bleeding red ink, that isn't particularly large in terms of sales yet (ie they're struggling to generate fast growth at a small'ish scale). And all that needs to support trading for 20 times sales on a business that is a decade old and has never produced a profit.
Either they find a lot of growth soon or in the next market down cycle Elastic is worth 1/3 to 1/2 of what they're trading for now. The same will probably go for their lesser peers. The clock is starting to tick hard on these extreme valuations (hello WeWork, Uber, Lyft).
Congrats to the team!
Although, am I the only one who finds that searching HN through Google gives better results than searching it through the Algolia powered HN search?
For example, search "ml" in both, Algolia results are years old and don't seem that relevant, whereas Google picks up more recent threads.
Their DocSearch program for open source projects is really great [0]. Check out the Jekyll docs [1] instant results that usually find just what you want. My only complaint is I want a plugin and play solution I can deploy for private internal documentation websites.
Congrats to the whole Algolia team.
I can’t think of any SaaS business that invested in, and executed such a smooth onboarding and retention ecosystem.
I’ve used them for small sites and large enterprise clients (10B+) and I’ve always felt like I’ve got way more than I’ve paid for.
PSA: Algolia has basically hidden a category making/taking over strategy in plain sight.
Hacker News generates 86.10% of referral traffic for algolia.com:
Amazing product and team. Nicolas still finds time to help small startups like us. Their success is well deserved.
Official blog post: https://blog.algolia.com/algolia-series-c-2019-funding/
Incase someone doesn't know already: they have a HN search too! https://hn.algolia.com/
What underlying tech does Algolia use? Elastic search?
I was just looking for a site search solution three days ago for a side project I'm working on. I found Elastic's offering but found the lack of a free tier(no 14-day trail crap) off-putting. I'll give these guys a try. Grats on the raise!
Does anyone know how MongoDB’s new search engine[0] compares to Algolia?
Congrats to them but I find the product to perform quite poorly on HN. Seems to do some kind of fuzzy matching / weird relevance ranking and not full indexing when all I really want is a full index text search.
Side question: how interesting and flexible is algolia as a replacement of a custom solr setup? I don't like the HN search and never heard that it is in use for larger data sets.
Are there any good open source alternatives?
If you've ever tried installing ElasticSearch and then switched to Algolia you'll understand how great of a product it is.
Now if Firebase could only buy them out and add decent search to their suite of products that would be swell. Mind boggling that Firebase - part of Google - still lacks a decent search solution.
There is a discussion on Stackshare(a bit old) about the tech they are using:
https://stackshare.io/posts/how-algolia-built-their-realtime...
Long time since those days in a tiny office in Rue du Sentier with Efounders...congrats folks!
I wonder if they can use the money to fix it so it works without Javascript?
Nice work.
The homepage is great from a developer perspective, select your backend on the left, frontend on the right, and you get an idea right away of what a basic implementation looks like. Very clever.
I figure with the ramped up marketing blitz they’ve been on we will be hearing about their IPO soon.
I’ve always had a hard time understanding the value proposition in the same way I don’t understand the value proposition of e.g. AWS Rekognition.
Paying per use certainly doesn’t make sense, because it has to be qualified by the accuracy you get per use.
And there’s no serious way to understand the accuracy you get per use (on your specific unusual distribution of queries) without employing the expensive ML / stats engineers you probably thought you could avoid hiring by outsourcing to Algolia / Rekognition in the first place.
But once you need to hire them anyway, you might as well utilize them to build this type of thing in-house in ways that are much more tailored and optimized around your in-house data models and data integration tools.
To put in perspective, I’ve worked in several companies (from small start-ups to large ecommerce sites) that have a variety of search needs spanning plug and play Lucene all the way to highly customized joint embedding neural network based nearest neighbor search, and tons in between.
The distribution of text in e.g. the support center search use case was totally different than the product search use case or the document store use case, where highly unique word distribution, special words, frequency of required updates to the search index, asymmetric costs of surfacing bad or deleted content items, etc., was the norm.
Every search use case was different and needed care to develop unique annotated result sets to measure mean reciprocal rank, NDCG, etc., as well as simple stakeholder subjective opinion of quality.
Short of basically hiring Algolia to be a gigantic consultant on all these things, I don’t see how it could actually be valuable.
I suspect it’s just an easy sell to CTO types that don’t really understand. They want “search” to be one problem with one little component to drop in to solve it, but it’s just not real.