SendPulse - Increase newsletter open rate with artificial intelligence.

@bogdan_su Bogdan, thank you for your question! In addition to traditional segments that are based on the data you have about users you can have segments that are based on user behavior. You can also combine traditional segments and behavioral segmentation and create your own very specific groups of subscribers..

@bogdan_su Thank you. Clearbit looks very interesting. Do you know other good sources of data about users? P.S. I checked out your project, very interesting idea. I had been an avid game player in the past and would love to check my stills :)

@shevkoplias You can also check out FullContact...they are similar to Clearbit

@vinitjoshi8 Thank you, FullContact is an interesting service! I will have a closer look at them.

@shevkoplias another one that I know is FullContact, but you might check this Quora thread https://www.quora.com/Are-there-... for some more P.S. Thanks for the kind words about Mobalytics=) What games you were playing?

@avinashdsouza Thank you for your comment and interest in our product. We discover user interests from the recent reactions to sent newsletters - how they read and click. Then we also try match what they do on the sites where we have subscription forms and tracking pixels. It's the first stage. On the next stage we plan to automate some parts of AB testing and content personalization. I'd love to hear your opinion, what user groups you think are important? What suggestions would you like to have - related to improving clickthrus, content or design? As for your question about the SMTP offer - 100K emails are "pay-as-you-go" plan and it does not include a dedicated IP.

@shevkoplias This looks like an awesome tool. I'm curious about the data you get from the subscription forms and tracking pixels. Would these sites that you track be interest-specific to the user?

@jenleeny The sites where we collect this data are the sites of the specific SendPulse customers. So, if you sell fashion wear, it's all about it, and not about say cars or electronics.

@manana_samuseva We are a global startup, based in US & Europe, and focused on Europe, Asia, Northern and South America ( absolute priority - BRICS, APAC and MINT countries). Therefore, SendPulse will add Italian, French, German, Turkish, Simple Chinese, Indonesian, Korean, Japanese, Arabic, and by the way, Spanish (LatAmerica), Brazilian Portuguese localization. Currently we've used pre-launch product for English and Russian audience. In total we will be have 13 foreign language groups. The language detection are based on geolocation. The next point about Artificial Intelligence. The modeling technology is based on math method, behavioral analyses to find lookalike audiences (digital twins) to create on behavior models in consuming content, clicks, design, social demographics - the simple comparison will be Facebook Artificial Intelligence for native advertising. Past interactions analysis makes lookalike (digital twins) audiences with the similar behavior patterns. As will as this is similar clones to brings more clicks to your site from the same newsletter. SendPulse Artificial Intelligence API is easy to use and does'nt require a complex setup process, and analyses hundreds of millions emails, dozens of millions daily pageviews, hundreds of millions events and in addition we also collect, buying own big data that can enrich personalization of the subscriber profiles. Using the big data of previous interactions, SendPulse makes recommendations how to improve the layout and the text of your newsletter before your send digital message (email/sms/push). Some A/B tests can be done automatically, like "communication channel test". So we reach customers the most preferred way. The algorithm will find lookalike audiences (digital twins) according to behavior patterns and will predict the best model to achieve best message delivery timing, geolocation via the best communication channel (E-mail, SMS, Web Push), and with mobile-friendly design and effective content. For example, if a user ignores some email but but reads and clicks on a similar SMS on his iPhone, then we should be able to set that as preferred delivery method, and during next campaign send him a SMS for the first step. So, the artificial intelligence is forecasts a model for effective frequency of contact with the target audience.

@manana_samuseva Thank you for your questions!

@savchukmark Thank you for your comment. Glad you find our solution interesting.

@savchukmark Thank you for your interest in our solution

@savchukmark You say Because they are not always "checking their email"... Totally agree! There's so called "first email bias". This is when you are more likely to read the top email in your list and ignore the rest. So, it's very important to deliver newsletter to the top of the list of unread emails. We try to predict when users are more likely to read email and respond. And deliver newsletters in the best time to give you better open/click rates

@manana_samuseva Thank you

@manana_samuseva Thank you for adding us to your collection! We have complex autoresponders that include emails, web push and SMS in one flow. Text messaging is international so you can send a message to any mobile phone worldwide (over 800 mobile networks in over 200 countries). We do have plans to support short codes sometime in future. Very few mobile networks allow MMS that are sent from web so we are not planning to add them. In most cases you can send a link to a file using a regular text message. Chatbots is an interesting direction, we will think about it. Thank you for your suggestion.

@manana_samuseva You seem to be very knowledgable about AI. I'd love to hear your opinion about chatbots. What bots do you think will be useful for a service like ours? Thank you!

@julia_eremina Julia, I really like your question! So, we have a secret sauce here. We analyze how your subscribers respond to your newsletter - if they open and click, what links they click, we do content analysis of the sent newsletter. It gives us invaluable information and we may predict how and when we can send better e-mail, it may increase the engagement. Our early tests show up to 50% increase to the email open conversion. So, the mission is possible!

@julia_eremina It's possible and we do that)

@julia_eremina The task is impossible if you just upload your list and want to predict the behavior looking into the addresses. It's impossible. By the way, after you send your newsletter several times and user respond (open, click), we will use this information. Also, if your set a tracking pixel to your site, it gives us even more information about your target audience interests. This all helps us to learn who are your users and what can we do to send more relevant traffic to your site from the same newsletter.

@arturkiulian Thank you, Arthur! We have started as a transitional email service provider and we have collected enough data about behavior of email recipients. We used that information to create user segments. When we do not have much information about the user, we allocate it to the segment(s) based on that small bits of information that we have.

@arturkiulian Our user profiles improve with every email you send to your user. The platform starts to understand what time you are most likely to read your messages, what time you are more likely to have a click through as well as a list of other presences we track. Our predictive modeling with learn and get better with every email and on the fly create the best scenario for a specific user when you want to send them an email and do it silently behind the scene. We would also love to hear your feedback on possible issues you may be having with your messaging and make recommendations.

@arturkiulian Hi Artur. About the cold start... Our approach is not to wait while the information is collected. So we ask you to send several newsletters to your entire mailing list. The actual number depends on your subscribers activity. So, at the first time when you create account, this functionality is not available. And it becomes active when you start working and sending. To make patterns, we use data from three main sources - collect it from your newsletter, collect it from your site if you set a tracking pixel, may buy data from such sources as FullContact, White Pages etc.

@paul_shuteyev Thank you stopping by, Paul! You have excellent open rates. We are in the process of creating the AI technology and it's just the beginning of the process. Any figure will be a wild guess. We do know that email open rates are going to increase but only tests will show. I hope you give SendPulse a try and take part in the testing and improvement process.

@paul_shuteyev Thank you for your question

@paul_shuteyev In average we see an increase up to 50%+ for open conversions. Would be happy to change your conversion from 20-30 to 50-70. Just imagine how more sales will you have ;)

@yaro How do you send it now? Using Rss-to-Email feature? To find out the best frequency need to try different time intervals and detect how users respond better - in daily, weekly or bi-weekly. It's exactly what AI should do for you.

@yaro The quality of your content will also play a big role in how often you should send content. Users can get overstimulated by daily emails if the content is repetitive or offers nothing new. With automated notifications it very important to send them at a time that users are most likely to need/want to read them. AI will help you automate these to avoid under or over communicating with your user. Have you manually attempted to test frequency of email?