I'm really happy where it's landed now and think the response has been much better this time around.

Smidge allows you to have quick and easy video conversations with customers from inside your app.

As a web user, I'm tired of seeing chatbots and chat widgets everywhere. If I want to read something, I'll read the docs. If I want to ask a question, I really just want to ask a person and get my answer as quick as possible.

As a founder, it's so hard to get customers into a conversation. We spend so much time deflecting them to KB articles and bots that we've lost the ability to engage with them 1:1 and that's what https://smidge.app is intended to do. We funnel users to the right team member and tee up a call instantly in-app.

Would love feedback whether video is a thing for folks (particularly early stage where you prob don't have a success team yet). We've found quick 5-minute conversations to be incredibly valuable.

(posted here 4mths ago with a different focus: https://news.ycombinator.com/item?id=24835602)

Comments URL: https://news.ycombinator.com/item?id=26322138

Points: 10

# Comments: 1

Also in the next couple of weeks, with just a user id, you will be able to place a call directly from Zendesk (for example) without us knowing anything about the user.

Down the track I believe we can deliver a self-hosted solution.

We don't show the user anything until you call them (so as to not interfere with drift/intercom etc). When you call them the current UI is like a mobile call UI.

EDIT: we are working on user initiated calling as we speak (it will show availability of support staff/CSRs)

Audio-only calling is in the next week or two

We had this problem with our product (BugHerd) and built a solution for it. Our success rate was ~1% and we thought we could do better.

We created a JS embed that you install in your web app. It allows you to segment your users and place a call with them from within your app (using webrtc). Unlike email/calendly/zoom square dance, Smidge has a 17-20% success rate to call. It's been huge for us and we're now opening to beta. Would love feedback

(Screensharing, "request a call" and page/user calling coming soon!)

https://smidge.app

As a founder, I’ve always found it hard to find customers willing to have a conversation about our products. The vast majority of users ignore chat apps and don’t respond to emails, but we keep using these tools because it’s cheap and easy to implement. It’s a habit, but it’s becoming less and less effective. I talked to dozens of other founders and found the same problems over and over again. We talk to customers during pre-sales and during support, but that’s about it. We talk to solve and sell, but rarely to learn. I wanted to do better. So I came up with Smidge.

Smidge is instant 1:1 video chat with users from within your app. No email, no Calendly links, nothing for your users to install. Just quick easy conversations.

You install our JS in your web app and we start keeping track of who is active in your app. When you want to talk to a customer, you simply click the “call” button and we robodial users one at a time until someone answers. It sounds intrusive, but in our experience, the users of BugHerd have loved it. It’s easy for them to decline or ask to be called another time (if they decline, we won’t ever call them again). If you optionally provide user meta data, you can create call groups to segment users into buckets. Helpful for getting your CSRs on to high value clients or helping people through onboarding.

We’ve been using it on BugHerd and the response has been pretty great. Previously we’d send out an email with a Calendly link, and out of 2000 odd emails, we’d get maybe 10-20 people on to a Zoom call. So < 1%. Smidge’s success rate is around 15-20%. That means within about 3 minutes of sitting down, you’re on a video call with a user. Until now that’s never been possible (unless you collect ph number at sign up). Our product team now typically talk to 1 customer each per day. That adds up to 250 customer conversations per year that simply weren’t happening before, and it requires no planning, just some spare time to make a call.

Very keen to get some more folks trying it, and of course I’d love your feedback. I’ll be watching here, but you can also email me at alan@smidge.app

Coming soon:

- recording/transcription to help you get insights into ProductBoard or Dovetail.

- Request a call with a specific user from a link (to allow you to hook that into Zendesk, intercom or other support apps).

- Page groups (so you can get CSRs to call folks on certain pages)

- and screen-sharing/remote-control.

Comments URL: https://news.ycombinator.com/item?id=24835602

Points: 3

# Comments: 1

It got very expensive for us to run, and we very nearly sold the app early on, but decided to keep hold of it and just add some little premium features to keep it afloat. Since then it's been running pretty close to break even. We kept it because we always had the intention of one day picking it up and getting serious about it.

Now that both Tenor (GIF Keyboard) and Giphy have sold to Google and Facebook respectively, we feel like it might be time to get serious.

What I'd love to know from the HN community is: a) do you have privacy concerns about Giphy/Facebook tracking? b) what platforms would we need to support to get you onboard with RightGIF c) are there app developers out there that would make use of our API if we made it public?

Thanks heaps!

Comments URL: https://news.ycombinator.com/item?id=23230938

Points: 34

# Comments: 19

At the time, we didn't entirely know whether there were still growth opportunities or what else we could try if we kept going. But we knew we wanted to keep trying. Some of our investors wanted out, and we gave them (and all the other investors) the opportunity to sell their shares back to us. It means everyone who wanted it got their money back (which for an investor is not actually a very good outcome, but better than $0), and we got to keep the company. I'm appreciative of them allowing us to do it.

Even though we were growing quickly, the data showed that we'd landed on a local maxima with the original product. Our churn rate was around 3%/mth and the vast majority of churn was "because we don't have a project in QA". We felt we could address that with a different offering (you'd call it a pivot). That pivot failed to deliver a product that resonated the way the original did.

Fortunately, we hadn't shut down the original product (which was still growing), and so the mistake was not fatal.