40+ SaaS Services Behind 15-Person Startup - 2nd Year in Review
Last year, we shared our detailed SaaS usage bills and triggered some discussions on HN.
This year, our business has grown multi-fold, and we further leverage the various SaaS services to maintain operational efficiency:
- We now own 3 open-source projects totaling 15k+ stars:
- We keep a lean team of under 15 members based in 4 different cities.
- We outsourced financial and legal services.
- As tool-makers, we are still picky about selecting the best tools for daily use.
As last year, below we list the SaaS services we use under 3 categories: Research & Development (R&D), Sales & Marketing (S&M), General Administration (GA). At the end of the post, you can also find the breakdown of the monthly cost and the comparison with last year's bill. Now, let's get started.
Research & Development
- (🌱 New) Airplane - internal tool
We migrated from Retool to Airplane to generate our enterprise license key. Internal tools are built by our developers, so Airplane's developer-oriented approach fits us.
- Algolia - document searching
- Auth0 - user sign-in
- (⬆️ Upgraded) AWS - service deployment
- Better Uptime - monitoring and alarming on service status
- Cloudflare - domain name
- (⬆️ Upgraded) Excalidraw Plus - prototype/sketch diagrams
All wireframing and product illustrations are done in Excalidraw. We subscribed to its team plan Excalidraw Plus as soon as it became available.
We announced Bytebase Cloud this year. The cloud service is built on GCP using its GKE and Cloud SQL service. Though we already used AWS before, we chose GCP because our founders used to work there and are more familiar with the platform.
(🌇 Sunset) Gitpod - cloud-based development environment
- Linear - project management
- Metabase - BI dashboard
- Neat - GitHub / Linear notifications
- (🌱 New) Neon - serverless database
We announced another project SQL Chat and it stores metadata and sample datasets in Neon. We tried a few other hosting options before. Neon's on-demand pricing model and its branching feature attracted us.
- (⬆️ Upgraded) Render - service deployment
We expand the footprint on Render. All service components other than Bytebase Cloud and SQL Chat are on render. This includes our demo environment, staging environment, Star History backend, SQL Chat backend and etc. The SQL Chat backend used to run on Vercel, however, its edge function is costly for the SQL Chat usage pattern. We have to ditch the edge function and adopt Render's pure server environment.
(🌇 Sunset) Retool - internal development tool
Replaced by Airplane mentioned above
- Segment - data integration
- Sourcegraph - code searching
- (🌱 New) Vantage - cloud cost management
- (⬆️ Upgraded) Vercel - website hosting
Sales & Marketing
Allow potential customers to self-schedule appointments with us.
We released our new marketing site early this year and reduced the 3rd party analytic services.
Star History now has a dedicated newsletter and we use Beehiiv as the newsletter platform.
- (🌱 New) Midjourney - image generation
We contract Midjourney to generate post images (not this one though).
Prepare product intro and investor deck.
- (⬆️ Upgraded) Plausible - user traffic analysis
- Paddle - payment collection
- Searchramen - SEO analysis
- (🌱 New) Stripe - payment collection
Bytebase Cloud uses Paddle because it takes care of Tax as a Merchant of Record. Stripe has been improving tax collection, so we decided to try Stripe and use it to collect payments for SQL Chat.
- Causal - financial analysis
- (🌱 New) ChatGPT - everything
- Google Workspace - Email and identity
- Grammarly - English writing assistant
- Lark - IM, documentation, office automation
- OSlash - shortcuts
- Pulley - equity management
- Slack - IM
- (🌱 New) Setapp - Mac app subscription
- (🌱 New) Zapier - workflow automation
Monthly Cost Breakdown
Research & Development
The spending increases by 250%, mostly from cloud usage consumption, Render, GCP, and AWS. This is a good thing to have. We are also using Vantage to monitor any cost cost anomalies. This month, we noticed an unexpected Snowflake usage from our Vantage report
It turns out it's a billing bug on the Snowflake side.
Sales & Marketing
The spending increased by 80%, mostly from increased usage (more email subscribers, having more visitors) and new sales and marketing tools.
Spending is decreased as we find ourselves not needing the paid features. We stopped our Grammarly subscription because we now use ChatGPT or other tools to refine the wording.
The total monthly spending is $2,227, a 90% increase from last year's $1,183!
|Research & Development||$425||$1,481||+250%|
|Sales & Marketing||$208||$368||+80%|
For startups like us, there is no Buy vs Build debate. Leveraging other SaaS services is a no-brainer. We also feel pain when there is no viable SaaS offering for a particular task. e.g. We still desperately need a service that allows us to write the post once and schedule it across all major social networks.
As with the rest of the industry, we are experiencing the AI storm. This year we have already onboarded a bunch of AI-native/augmented SaaS services. We even built one, SQL Chat, offering a natural language interface to the database. I am already used to casting spells to construct the Zapier workflow or co-author pictures with Midjourney.
Compared with last year, we are using more SaaS services as we grow our business. It's still manageable.
Feel free to share this article, and if you are using some neat tools not mentioned here, we'd love to hear about them. See you next year!