Battle Cards

  • Algolia

    Overview

    Search is a competitive advantage for giant sites like Pinterest, Amazon, and Netflix. Algolia is search for app developers who see search as a necessity, rather than a competitive advantage. If you actually want to build best-in-class search into your app using Algolia, be prepared to pay for it. The more advanced and secure your search engine is, the more it ends up costing.

    Algolia is:

    1. Fast
    2. Easy to use, and 3.) Easy to integrate.

    Algolia would honestly be more appealing had it started ​before​ Elasticsearch. Now, they’re just having to play catch up to a very vibrant open source project. Algolia touts that they are a “search platform,” but they aren’t even scratching the surface ​in terms of comparative popularity to Elasticsearch. They easily integrate with tools like Magento, Woo Commerce, WordPress, Heroku and more. While this may seem impressive, ​Elasticsearch integrates with more tools​. It has more advanced features, and will continue to gain more as the market matures. Any customer who chooses Algolia is locking themselves into a search engine than competes with the open source technology used by Etsy, Pinterest, Netflix, Spotify, and loads of other tech companies. Hmm…

    Algolia has a reputation for being expensive, which they recently combatted by ​releasing a version for $479/mo​. This offers more features than their starter plan, and honestly, you can’t really build a decent search engine without it. This plan gives you decent query rules that allow you to do stuff like boost certain results ( = free in Elasticsearch) based on IP address or profit margin.​ ​These are fairly standard things that search-based sites do. They give you control over how search results are sorted. You also get access to their API, which is what you’ll need if you want to do Kibana-style ( = free) reporting.

    So now into personalization and AI-based search, which is the secret sauce a lot of the cool kid tech companies are using to differentiate their products. This costs no less than $3500 a month, and requires an annual contract. You need this plan to build complex drill down search that allows for ​broader faceting​ ( = free in Elasticsearch with complex aggregations), or even serve results based on search history, account characteristics, or geography. Personalization sounds advanced, but it’s really not when you think about how prevalent it is. At that price, our team will actually hop in your Slack channel or on a call, and help you with Elasticsearch. That’s $42,000 a year. We will help you with architecture, and you’ll likely still have some money left over for custom development.

    Not only is the product itself expensive, Algolia forces you into the “Contact Us” tier just to get dedicated infrastructure and enterprise support. So safe to say, you will pay over $42,000 a year if you care about enterprise hosting/support that you can get from Bonsai for $14,000 a year.

    This is particularly important in the banking and healthcare sectors, who will likely require dedicated infrastructure but may not want to pay this much simply for search.

    Summary

    If your team:

    1. Doesn’t see search as a competitive advantage you are willing to build,
    2. Is willing to pay at least $6000 a year for decent search, and $42,000 for dedicated and/or personalized or advanced search,
    3. Will take straightforward documentation/ease of use over more capabilities,
    4. Doesn’t want to even think about infrastructure or sharding,
    5. Thinks a proprietary search engine that is a few years old can compete with a vibrant open source technology that is ten years old,

    Algolia is for you. If not, stay away.

    Have a look at Algolia’s detailed feature set at the ​bottom of this page​.

    Key Phrases to Keep Customers Off of Algolia:

    “What features are you looking to incorporate in your search engine?” → find equivalent in ES “Do you foresee your team planning to customize or test your search configuration?” “Have you looked at how much those features cost in Algolia’s pricing model?”

    Faceoff

    Strengths

    1. The easiest way to get powerful functionality out of search.
    2. Awesome, easy to read documentation.
    3. Completely hosted. Customers don’t even need to understand sharding.

    Weaknesses

    1. Proprietary and not as robust as Elasticsearch.
    2. You basically have to sell your firstborn just to get the functionality that Elasticsearch offers for free.

    Positioning

    Turnkey powerful search

    Repositioning

    Bonsai takes the pain out of Elasticsearch. Why pay for a proprietary search engine when the open source version is better?

    Pricing Comparison

    Not really an apples to apples comparison. Basically, you pay not to have to learn anything about search.

    Their Target Market

    Mostly front end developers. Algolia is popular in the JavaScript community.

    Market Presence

    Gaining traction, especially among front end developers.

    Quick Tips

    Point out how much you have to spend just to get basic stuff already in Elasticsearch. Algolia is slick, but a simple Google search of “Elasticsearch (feature)” shows that Elasticsearch does everything it does, and much more.

    How to Win

    Do the math. You can get the basics out of Bonsai for much less. You can also get advanced features with the help of our team, and still save money. Have Dan, Dru, or Rob hop on a call to highlight Elasticsearch’s capabilities.

    When to Walk away

    When a customer sees search as a pain to solve rather than a competitive advantage, and doesn’t care about the lost opportunities of choosing a much more robust technology.

  • AWS Elastic

    Overview

    Amazon released their own version of Elasticsearch in 2015. While they were ​criticized early on for not supporting features like node-level metrics and VPC, they have since added these features.

    AWS Elasticsearch is not ideal for a search or DevOps newb who “doesn’t know what he doesn’t know.” AWS does not set up Dual DNS out of the box, and they don’t have security built in. For this reason, ​thousands of AWS users’ Elasticsearch clusters​ were hacked in a ransomware attack. They support metrics via Cloudwatch, but you ​have to know what to pull in from their API​. You have to know ​what instance types to pick out​ when you sign up. This is not a task for the faint of heart. AWS also forces you to pick all of your instance types, security, and AZ settings upfront as soon as you sign up. It’s like they’re expecting that you’re moving from some other service.

    Out of the box, AWS does not come with support. Customers can add their ​own support packages based on their spend​. The more you spend, the faster their incident response is, and the more access you receive to Elasticsearch experts. AWS has a tiered support system. Your ticket is only escalated to an expert if required. You can get ​Trusted Advisor​ checks, but these are not Elasticsearch experts — these are AWS experts. How can they help you with shard counts? Will they understand the Elasticsearch features that will make your index more efficient? Unlike Bonsai, you will not receive any guidance at all unless you pay for a business tier support package, which is $100 a month. You don’t get onboarding. You will not get enterprise support unless you spend $15,000 a month on AWS.

    Summary

    AWS is good for the “Search Expert” who has a deep familiarity with Elasticsearch, distributed systems, and AWS. It is also good for giant customers who can afford expensive AWS support contracts. It could be a liability for developers who are “Search Pragmatists,” who would likely not want to spend their time securing, managing, and scaling Elasticsearch. Elasticsearch is a massively distributed system — it’s best to leave the hosting to the experts!

    Key Phrases to Keep Customers Off of AWS

    “Sure, you can save some money moving to AWS, but you’ll have to spend a lot of time learning and managing Elasticsearch instead of building your app/company.” “Bonsai has been managing search engines since 2009. Elasticsearch is an afterthought for AWS. You literally have to spend $15,000 a month before you get the expertise we’ll give you for $600 a month.”

    Faceoff

    Strengths

    1. Integrates with Amazon products like Cloudwatch and Kinesis.
    2. Pricing is highly configurable based on need. Customers can pay anywhere from $13 to hundreds of thousands per month

    Weaknesses

    1. Not easy to use for developers who are not experts at Elasticsearch and AWS.
    2. Support tickets are not immediately sent to ES experts like they would at Bonsai and Elastic.

    Positioning

    Elasticsearch that plays well with other services on AWS.

    Repositioning

    Elasticsearch managed by people who are not ES experts.

    Pricing Comparison

    AWS ES is always cheaper than we are. As a customers’ search cluster grows, this cost difference magnifies. To use AWS ES though, you have to have a DevOps and an Elasticsearch Engineer (= $100k+ a year).

    Their Target Market

    AWS power users

    Market Presence

    “The elephant in the room.” Their Elasticsearch product gets more ​search volume on Google​ than both us and Elastic combined.

    Quick Tips

    If a customer has many years of experience with AWS AND ES, we’ll likely lose the deal. If he or she has built massively distributed systems, he or she will likely be okay with AWS ES. If not, engage Dru or Rob to kindly test the customers’ knowledge and confidence.

    How to Win

    1. Find out how familiar a customer is with AWS and Elasticsearch.
    2. If they are not familiar with both, steer them away from AWS Elasticsearch.

    When to Walk away

    When a customer is comfortable with AWS, AND he or she is knowledgeable about Elasticsearch best practices, or is very willing and has the time to get there.

  • Elastic On Prem

    Overview

    Elastic is very big on selling not just Elasticsearch, but “The Elastic Stack.” The Elastic Stack is composed of:

    • Elasticsearch
    • Logstash
    • Kibana
    • X-Pack

    The ELK components of this package are free and open source. It is only X-Pack that must be purchased. So when we sell against this solution, we are selling 1.) our service against their service 2.) our features against X-Pack, 3.) our uptime and performance and 4.) price.

    So how do we stack up?

    For customers who are fully committed to Elasticsearch, on-prem Elastic is a good bet. But they must be fully invested. They have to be committed to setting it up and maintaining it. They have to know how to scale it. Elastic has ​classes for all of this​ that come with a hefty price tag ($400 for a single course), or, a company can purchase an annual subscription.

    Elastic doesn’t sell just support. They sell support with X-Pack (features listed ​here​). This is $4400 per node for gold, and $6600 for platinum. This isn’t just incidence response (although that is a part of it) — Elastic is offering “Ask Me Anything” kind of support for this price. Customers can choose X-Pack with our dedicated option, but they would be purchasing support twice. An entire article about X-Pack is coming soon.

    Summary

    If a customer is ready to dedicate resources to hosting Elasticsearch, AND has the money/time to familiarize themselves with X-Pack’s features, AND has the money to pay this kind of pricing for support, Elastic is a great option. Enterprises that are willing to throw down big cash for “big data capabilities” will be hard sells. An Elastic customer has to be “full in” though. This option takes a lot of time/money.

    Key Phrases to Keep Customers Off of Elastic On-Prem

    “What do you need in X-Pack?” (we may have comparable features that they just don’t know about. “Is your DevOps team prepared to host Elasticsearch?” “How familiar is your team with Elasticsearch or distributed systems?”

    Faceoff

    Strengths

    1. Support from the makers of Elasticsearch.
    2. X-Pack gives customers a variety of useful tools for Elasticsearch.

    Weaknesses

    1. Expensive. $4400 per node on a production cluster is over $13k a year.
    2. A lot to set up, and a lot to learn. Requires dedicated Ops and management resources

    Positioning

    Best in class features/support from the makers of Elasticsearch.

    Repositioning

    It’s good, but only if you have the time and resources to fully invest into Elasticsearch.

    Pricing Comparison

    Not really an apples to apples comparison. We manage servers for you, but don’t offer all the features in X-Pack.

    Their Target Market

    Enterprise, enterprise, enterprise!

    Market Presence

    Goes without saying.

    Quick Tips

    Figure out how much bandwidth they have to host their own instance of Elasticsearch, as well as which features they need in X-Pack.

    How to Win

    1. Sell the features that we have that are similar to X-Pack.
    2. Price. If they aren’t committed fully, they’ll be wasting their money.

    When to Walk away

    When a customer seems fully committed to investing budget and headcount for Elasticsearch.

  • Swiftype

    Overview

    Swiftype​ started as a SaaS service on top of Elasticsearch. It essentially obfuscates the complexity of Elasticsearch. Elastic ended up buying their service. They have two key offerings:

    1. Site Search, which is essentially what Google Site Search was.
    2. App Search, which competes with us.

    Elastic packages it all up under the name “​Site Search​” on their site. They are two very different products with different use cases. We’ll focus on App Search.

    App search would be likely be what Bonsai customers would look at here. Users get a simple UI to load in their index, create synonyms, facets, etc. It is an experience that is custom catered for app developers who want to use key features in Elasticsearch, but don’t want to have to learn anything about infrastructure. Unlike Bonsai, who charges mostly on metrics like memory and storage, Swiftype charges by documents and API calls. This means if you have a few items and not a lot of traffic on a site that will likely not be around long, you’re set. You can see that they have ​three primary pricing tiers​ — Standard ($49), Pro ($199), and Premium ($1999). It can be very expensive if you need more than one search engine, e.g. one for customer records, and one for inventory, or one for staging and one for production.

    Swiftype’s ease of use comes with a cost though, and this is where it helps to embrace pure open source technologies like the ones we sell. Bonsai users can use one-click Kibana for free. Swiftype users only get monthly Analytics on their $199/mo plan. It’s also not ideal for many use cases. You likely would not use Swiftype for log retention, or anything other than basic vanilla consumer app search. It’s not customizable or ideal for scenarios with a lot of writes to an index, e.g. user generated content. It doesn’t support many languages like Elasticsearch does.

    Swiftype handcuffs users to their proprietary app and pricing model. You can’t duplicate and stage multiple cluster configurations for testing, which is easy to do on Bonsai. Well, unless you buy an entirely separate plan. If you get excited about some cool feature in Elasticsearch, guess what? Not only can you not customize it — you don’t even get it. Swiftype is proprietary. You get a few of the key features of Elasticsearch that are helpful for a specific use case. That’s it.

    Summary

    If your use case is:

    1. A basic web or mobile app with one search engine, and records you load yourself, and
    2. There’s likely no way you will ever need to customize or test your search beyond the basic synonyms/ranking/facets stuff, and
    3. You are okay spending money on features that are available for free in Elasticsearch,
    4. You do not care about many other features Elasticsearch offers,
    5. You do not want to bother with any elements of hosting your search engine,

    Swiftype is for you. Otherwise, it’s best to choose an Elasticsearch option.

    Key Phrases to Keep Customers Off of Elastic Swiftype

    “How long do you plan to keep this app running? How important is this search engine for your app?” “Do you foresee your team planning to customize or test your search configuration?” “Does your app require a lot of writes to your search engine that might require some configuration?”

    Faceoff

    Strengths

    1. The easiest search engine to use on the market.
    2. Custom catered for apps that only need one search engine.
    3. Completely hosted. Customers don’t even need to understand sharding.

    Weaknesses

    1. Not nearly as customizable or extensible as Elasticsearch.
    2. Proprietary and not as robust as Elasticsearch.
    3. Expensive.

    Positioning

    The power of Elasticsearch for app developers, without the complexity.

    Repositioning

    A limited version of Elasticsearch that many customers will likely outgrow.

    Pricing Comparison

    Not really an apples to apples comparison.

    Their Target Market

    App developers who have no time to learn anything about Elasticsearch.

    Market Presence

    In the Heroku marketplace. Just got a little more prominence on elastic.co.

    Quick Tips

    Get a sense of “where the customer sits in the stack.” If they are front-end developers who are scared of databases in general, they’ll likely gravitate towards something like Swiftype.

    How to Win

    1. Send them the ​Ideal Elasticsearch Index​ blog post, and let them know we can help them.
    2. Remind customers of what they miss out on by choosing Swiftype.

    When to Walk away

    When a customer has zero interest in learning about Elasticsearch, and has the money and use case that makes that possible.