[Finally launched in beta] Aiming 100% uptick in customer engagement with personalization.services
Hey everyone,
It took a while pondering over the ideas coming up with the right technology and product value. Finally it's here - https://personalization.services
What does this do?
Offers set of content personalization REST APIs for ranking and recommendation use-cases, backed by machine learning technologies.
What's new or different?
-
Hands off the wheel: The Uptick technology doesn't require any data ingestion or explicit model training. The API integration automatically detects newer entities and changing data and adapts to that. So the only effort is one-time API integration.
-
Super fast recommendations: The similarities recommendations API is able to generate super fast recommendations which can directly be integrated on front-end or backend for real-time use-cases.
-
Cohort analysis and customer segmentations - Personalization can sometimes appear creepy. Uptick personalization engine conducts cohort analysis in order to find out similar set of customers and recommendations for them.
Lessons so far
- Speed matters in online business. Even extra 300ms can have engagement effect. That was the motivation for keeping the API latencies low.
- I've been an angular + AWS pro, but I think when it comes to creating landing websites, nothing beats the agility of wordpress (with elementor). I learnt that separating the landing websites from more technical and specialized dashboards (written in angular etc.) is actually a great idea.
- I tried getting some early feedback by talking 1:1 with people on anonymous chat website Omegle. Didn't work out quite well, as only 1 person checked it, most won't even bother, and that one person wasn't a target person.
Next steps
- Work very closely with early clients and continue to evolve backend infrastructure.
- Continue on deep personalization with industry experts and scientists.
What I'm looking for
- Feedback. Seriously thrash it if needed but please, please, say it all, or just say Hi. Your feedback is very important.
- Early adopters - Over the next few months, I'm willing to work very closely with our early adopters in order to solve some of the customer problems for them. So if you're one, please reach out - https://personalization.services/contact-us/ or at [email protected]
- Mentor - I'm also looking for a mentor to guide me on market adoption and launch strategies. If you're willing to help, please let me know, I'd really appreciate it.
Also, feel free to AMA (ask-me-anything).
Congrats Sanjay, I can tell you've put a lot of work into this. You asked for constructive feedback, so here is what comes to mind:
Is the orange "play" button icon on the hero graphic going to be clickable later? I ask because it's the same color as your CTA buttons but not clickable.
Your four icons (powerful, intelligence, scalable, automatic) looks great, the designs are tasteful and unique. The scroll over color change effect on the icons is nice but not functional, which is a slight detractor for me.
The measure growth section looks and feels a bit empty. The analytics laptop graphic on the left looks like stock imagery and the text is fairly simple. Consider adding a second sentence to the subheader, or even better, putting something clickable here that allows people learn more about your growth measurement tools.
The "ready to delight" and "consult your use case" sections are very similar in their visual design, which create a feeling of redundancy. They also don't have a lot of content. You could consolidate them into a single section with two CTA buttons.
Hi Ezra,
First of all thanks a lot for writing detailed feedback on this. I really appreciate it and working on incorporating all of it. I'll update this comment as I do.
Again thank you for taking time for writing this, really appreciate it. - Sanjay
Great, I like the changes you made. Looks much better now!
Sounds interesting. I'm curious as to how you can make suggestions based on what you know about a user. What kind of data is a necessary requisite for getting useful information out of the api?
You're in seattle? Someone should schedule an IH meetup in this awesome city. I'm based here also. :)
Good luck.
Hi Nikk - The recommendations lie in how similar customers are, in their behavior. And the similarity is determined by what kind of recommendations they engage with, along with other behavior and attributes. More and diverse data about a customer is always better and helps represent them more correctly, however there is no hard requirement and no necessary data, as the recommendations platform is designed to work with any arbitrary data. Even in the absence of any concrete data, it can rank just based on IP address / geography, time of the day, device type etc., ultimately recommending what most people engage with in these attributes (just a basic example for illustration). Even in the absence of those, it can still recommend what most people are most likely to engage with. The more specific the data about customer and context, better the ranking and recommendations would be. I hope that makes sense - feel free to ask more.
Yes I'm in Seattle and I love it, +1 on the IH meetup idea, would love to join one.
This is really interesting I think it is really important to remember that online personalization can take your website from good to great. I recommend the site twik for all things personalization.