Ask IH: How do you handle SaaS availability when starting?
Hi, we're about to launch our new product as a SaaS solution and we are trying to find out how we are going to be able to handle the 24/7 availability of a SaaS product in terms of chat/mail support but also in terms of service availability when not working and on weekends and holidays.
Any experience/advice?
Glibly, don't make a service that goes down and you'll find that you won't need much in the way of live support. If your business model really does requires round-the-click human support (high-touch, high-value transactions), hopefully these costs are baked into your pricing already.
the product has been running without problems in-house for over a year.
regarding human support, we just don't know how much and of what kind we are going to get, we are doing a soft launch now with known people to get an idea of it and later do a public launch.
I can do some support from my phone for the chat and emails, the thing that may be a problem is if I'm offline or away from a computer and something happens on our service that needs manual intervention.
There is a managed service with tawk.to that they charge $1 an hour for live support.
Also if you feel uncomfortable create some loud alerts that can wake you up if sh*t hits the fan till you reach a point you can have support around the clock.
I'm mostly scared about getting the alert and being far from a computer or without internet access.
You learn to not be far from a computer or without internet access. That's basically what being on call is. When on call my laptop goes to the shops with me, it goes to the beach, it gets to go hiking(where there is phone reception), etc.
Do your service demands live chat? (It is live support oriented?) If not, then you worry for nothing. Having simply "We are offline now, our online hours are X-XX. Fill out the form and we will reply to you as soon as we are back!"
live chat is mostly for people on the landing page
You worry too much. Downtime for an hour or two is not an end of the world once in a while. Literally, I think you should refocus on your product instead.
Unless your product won't work if it won't be online 99.9% of the time (like uptime checker or financial tracking tool, where every second count) then you need to stop fighting windmills.
I would suggest that you carefully examine whether you absolutely need for-real 24/7 live support.
My bank offers a phone app to check my balance and transfer money between accounts. These are important tasks, and yet the app is frequently “unavailable” for mysterious reasons. I am still a happy customer of this bank. It would be cool if the app was always available, but it’s an imperfect world.
As you’be seen from the excellent replies in this thread, if you really do need constant support then you’re going to have to hire people. That’s expensive and adds complexity to your business. If it were me, I would only offer that kind of support if the customers demanding it were very high-value and would churn out otherwise. I think most people are fine with sending an email and getting a response back in a timely manner, especially for problems which aren’t matters of life and death.
Obviously, if someone could literally die if your SaaS goes down over a long weekend this advice doesn’t apply to you 😄
We're a fairly small team here at Palar (https://palar.support), but we're able to provide support and availability 24/7/365 to our largest customers.
Here are some of our tricks:
Have a remote CS team.
Subtract 12 hours from your timezone, and find someone there.
It's a start.
You could work around that but eventually it will be inevitable or simply impractical to not have one.
Organize and be realistic.
Do you have different levels of support and SLAs? Organize them in your ticketing and chat systems.
We would all love to provide 10 minutes responses to all customers, but how realistic would that be? Does it make sense to provide lightning fast responses outside business hours to users on Free plans?
A simple approach is to only prioritize your highest paying customers. If someone is providing a large share of my revenue, I'm more than happy to wake up at 3am to answer them.
3)Processes. Processes. Processes.
It's cliche, but it's essential.
You want to plan in advance of any incidents or emergencies.
Build reasonable escalation policies around your team making use of a service like Pagerduty.
Make sure someone responsible for each part of your platform is reachable at anytime you might need them.
A number of great live chat products like Drift or Intercom have the concept of “office hours” so the chat automatically conveys to the user if humans are online or not.
You can also integrate your knowledge base with these chat solutions so the chat automatically recommends help articles while the person waits.
Lastly, Drift allows you to build out chatbots to script more complex conversation flows with users.
It’s not uncommon to automate 65-80% of conversations on your marketing site with a bot.
I’m not professionally affiliated with either—just help businesses generate leads through live chat on a consulting basis.
Let me know if I can clarify anything.
Best of luck!
I've been dealing with the same question with the launch of my service clicktoverify.net. What I settled on was: I am on call 24/7 for the app breaking (I have some monitoring in place that will ring my phone, even if I'm asleep), but I'm not on call 24/7 for support of a specific customer (integration issues, etc).
I do have a welcome email that gives new signups an "urgent" support email address, but I don't set the expectation that it will be answered immediately, only that I will prioritize urgent issues more than others.
Now, in this period where I'm launching and trying to totally delight every signup, I am in practice on 24/7 call. My phone is set to alert me when email comes in to the "urgent" address, and I do wake up to answer these emails. This sucks, but I think it's worth it at early stages while I'm trying to build a relationship with my early customers and understand them. I'm on call for my day job anyway, usually for issues I care much less about than gaining a new customer, so it's not much additional hardship.
In other words, be honest about who you are (don't let people think you're a big conglomerate with 14 offices on 12 1/2 continents, when you're just a guy with a day job and a laptop), and most importantly underpromise and overdeliver. If you say "we try to respond within the day" and always answer within 5m, you've wowed somebody. If your site says "99.9% of calls answered in 2 minutes" but you flub one and miss it, you seem like a huckster.
Hi and congrats.
For support have 2 options. The first (and this is my favourite) is to hire two support people in the Americas, two in Europe and two in Asia who work 8 hour shifts and cover the clock and weekends between each other.
The second which is what my company did previously but I find it inhumane and also bad for employee morale and retention is to hire a team in one location that works around the clock on shifts. Our team was in Philippines at the time.
In terms of app availbility people expect 99.99% uptime.
6 hires for support is a lot for us at this point, I can cover the european time zone and my co-founder the american one, we will see how much timezones are left to cover and maybe hire someone for that shift.
Obviously 6 can be 3 but then you are left with no weekends and covering if someone is ill or on holidays. An alternative is 3 + 3 part time.
That's totally up to you.
The important thing is to be upfront, honest and consistent.
If you think that your customers wouldn't sign up for your service if they knew that it might be down (or without support) occasionally on weekends, then IMO you need to provide support/uptime guarantees on weekends. Expensive/difficult? Then raise your prices and plan/hire accordingly.
Otherwise just make sure to write into your TOS that you won't be providing support/guaranteeing uptime at certain times and on weekends. And make sure your customers are aware of this when they sign up.
Bad things happen when you promise the first but deliver the second.
Good luck!