This month marked the end of the first quarter of the year in a particularly exciting way for digital marketing enthusiasts.
The digital marketing news and updates during March were exciting, especially in the search and AI industry.
An AI clone named Dolly – yes, like the cloned sheep – has appeared to democratise large language models. On the other hand, ChatGPT-4 has launched, positioning itself as a threat to Google with its multimodal model, and Google has announced a new core algorithm update to spice things up in the search field.
The marketing technology field has seen its share of updates with the introduction of new features for two of the most relevant tools for marketers and managers: HubSpot and Salesforce.
These updates are exciting, and to stay ahead of the game, you’ll want to read the full article.
Source: Pexels – PhotoMIX Company
Google has announced the start of its March 2023 core algorithm update rollout, and naturally marks the beginning of search marketing this year. We understand that these updates can have a major effect on search rankings, which is why staying informed is essential.
Please keep in mind that core updates may take several weeks to fully implement, and search rankings could change over that period. It will be impossible to determine the full impact of the March 2023 core update until the rollout is closer to its end. So we can predict this will be an ongoing topic.
What do we suggest having in mind with this update?
Google is well-known for its algorithm updates, which it introduces regularly to improve its understanding and ranking of content.
These updates aim to provide increasingly relevant and high-quality search results. However, they also bring changes to how Google evaluates and ranks websites. As a result, SEO professionals and website owners must stay informed and adapt their strategies accordingly.
As we await more information on Google’s March 2023 core algorithm update, there are a few key takeaways to keep in mind:
- Allow time to work its magic: it can take days or even weeks for Google’s crawlers to fully assess the web.
- Measure organic traffic: Make sure to monitor your site’s organic search traffic and keyword rankings. These metrics can help you determine the impact of the update and identify potential areas for improvement.
- Don’t write for the rankings: Delivering quality content that is tailored to your target audience’s needs is essential. The goal of these updates is to provide the best search results for your users.
- Optimise your site for rankings: Technical aspects such as site speed, mobile-friendliness, and proper indexing have a significant role in determining your overall ranking, so pay attention to the fine details.
On March 14th, the new multimodal version of OpenAI GPT was launched. It’s safe to say that if Google wasn’t already wary of this popular AI tool, they have even more reason to be now.
What does it mean to OpenAI CPT-4 to integrate a multimodal model?
Multimodal AI refers to the ability to process various types of input, such as video, images, and sound. In machine learning, the term “modality” refers to the different forms of input, including text and senses like sound, vision, and smell.
As GPT-4 states:
“We’ve created GPT-4, the latest milestone in OpenAI’s effort in scaling up deep learning. GPT-4 is a large multimodal model (accepting image and text inputs, emitting text outputs) that, while less capable than humans in many real-world scenarios, exhibits human-level performance on various professional and academic benchmarks”
What are some of the improved capabilities of GPT-4?
- GPT-4 is more reliable and creative than GPT-3.5
- GPT-4 is able to handle much more nuanced instructions than GPT-3.5
- The distinction between GPT-3.5 and GPT-4 is subtle in casual conversation, but becomes apparent when the complexity of the task reaches a sufficient threshold
- GPT-4 can handle text and image inputs, which means users can assign any language or vision task they want, not just text-only tasks.
- GPT-4 appears to work across all languages.
Below, you can find an example of what adding this visual inputs to OpenAI GPT-4 can bring to the table:
Undoubtedly, this is where GPT-4 steps stronger, thanks to its exceptional ability to incorporate inputs in a more natural language through the integration of this multimodal model. Over the last year, we saw Google striving to implement a more human-like Search in 2022, using different tools such as Google Lens, etc.
3. Introducing Dolly: A New Open-Source ChatGPT Clone
Dolly and clones are two words that historically have pared well together, and in this case is no different.
Databricks, an enterprise software company, recently released a new ChatGPT clone called Dolly. The clone is named after the famous sheep that was the first mammal to be cloned. The Dolly Large Language Model (DLL) is another step forward for the open-source GPT Chat.
Open Source Large Language Models in AI Development
The Dolly Large Language Model (DLL) is the latest addition to the open source AI movement, which aims to provide greater access to technology so that it is not monopolized and controlled by large corporations.
One of the driving concerns behind this movement is that businesses may be hesitant to hand over sensitive data to a third party that controls the AI technology.
The Open Source Model
Dolly was created from an open-source model developed by the non-profit EleutherAI research institute and Stanford University’s Alpaca model, which was in turn created from the 65 billion parameter open-source LLaMA model developed by Meta.
LLaMA, which stands for Large Language Model Meta AI, is a language model trained on publicly available data. According to a report by Weights & Biases, LLaMA can outperform many top language models, such as OpenAI’s GPT-3, Gopher by DeepMind, and Chinchilla by DeepMind, despite being smaller.
What happens with an Open Source model for AI?
Open source in AI offers several advantages, including cost-effectiveness and the ability to customize to fit a company’s specific needs. Collaboration between developers is encouraged, leading to better AI solutions. Innovation and experimentation are promoted, resulting in better AI algorithms and solutions. Transparency is also provided, allowing for better scrutiny of algorithms and code.
However, open source AI also has its disadvantages. Security vulnerabilities and malicious attacks can occur, and the quality of code and algorithms may be lower than that of commercial AI developers. Ongoing maintenance is necessary, and fragmentation can occur if different teams and developers are working on different versions of the same project. Inaccuracies may also arise with open source AI.
Comparing ChatGPT and Dolly, databricks mentioned:
“We evaluated Dolly on the instruction-following capabilities described in the InstructGPT paper that ChatGPT is based on and found that it exhibits many of the same qualitative capabilities, including text generation, brainstorming and open Q&A. Of particular note in these examples is not the quality of the generated text, but rather the vast improvement in instruction-following capability that results from fine-tuning a years-old open source model on a small, high quality dataset.”
Source: Deposit Photos
Salesforce, the leading cloud-based CRM platform, has been helping businesses grow and succeed for many years by organizing and managing customer data. With over 100 new features and updates released, Salesforce continues to be the top choice for businesses of all sizes.
Recently, Salesforce rolled out its spring update, which included a number of exciting new features. In particular, we were impressed by the following:
- Use Lightning Web Runtime (LWR) to build faster and cheaper websites:
The B2B LWR template enables easier customization of storefronts. Additionally, the improved functionality and performance of the framework results in faster run times and more personalized experiences for customers.
- Salesforce Payments now works more seamlessly:
Quick payment processing is essential for ecommerce businesses. Salesforce’s new payment framework makes the process smoother and requires less backend coding. It is more user-friendly and allows for third-party payment providers to be integrated into the checkout process. A new click-and-go function streamlines the checkout process for customers.
- Salesforce CMS now supports approvals, viewing, and restoring including: the ability for consultants and contributors to create, edit, and view content in their workspace, and for other users to approve or request changes. Additionally, users can now view and restore previous versions of content, track changes, and revert to older versions if needed.
- Metrics and analytics:
Sales Analytics now has better filters and quicker access to data. This lets you find what you require without going through unnecessary data. Sales managers and reps can see which sales strategies work best, which products are doing well, and find things to improve.
- The Account Engagement Optimizer can assist you in:
- Identify ways to improve your marketing automation system, such as optimizing sales processes and email campaigns, to increase ROI.
- Identify risks to your marketing automation system to improve overall marketing. For example, the Account Engagement Optimizer may flag campaigns where click-through rates are below industry averages and prompt changes.
- Prioritize recommended changes based on what’s most important. These suggestions will help your team focus on changes with the biggest impact.
Source: Deposit Photos
If you are a business manager, it is likely that you are a HubSpot user. The tool has been publishing more comprehensive monthly updates outside individual portals, such as the “What’s New in HubSpot’s Software” page. We are here to tell you about the 5 latest updates that might come in handy.
- Marketing Campaign custom properties (beta):
Organise your campaign information and assets more efficiently by creating custom properties for Marketing Campaigns. The once-limited Marketing Campaigns feature now behaves like other standard CRM objects (i.e., contacts and deals), eliminating the need for complex naming conventions to store campaign information in the name.
- Check bot filtering status to track campaign engagement more accurately:
Check if your email engagement metrics are accurate by verifying if bots are being filtered out. Look for the new Bot Filtering setting on the email post-send page. If your email engagement metrics are different after migrating to HubSpot, toggle this switch on and off to see if bots were the reason for the higher engagement numbers in your previous system.
- List releases: Create a random sample, clone and restore static lists, and filter releases based on ‘has/has not completed’:
If you conduct A/B testing, personalise actions for specific lists, or randomly select contest winners, the new beta release enables you to create a random sample of any list. This feature eliminates the need for previous workarounds and potential errors from exporting and importing spreadsheets, and allows you to choose a percentage or count of list members to add to a new static list.
- Coloured deal tags (beta):
Use the new beta feature for coloured deal tags to prioritize your constantly changing deal pipeline boards.
- Easier social media publishing and email replies on mobile (beta):
The new HubSpot mobile app features help your team create and edit social media posts on the go. Additionally, the new beta release lets you send and reply to emails within the app. This ensures timely, tracked emails and enables all team members to see the latest conversation and context for the record, regardless of the device used.
Source: Deposit Photos
We can say that 2023 has been an eventful year for TikTok. In January, due to data security & privacy, the company was already facing legal consequences in the EU.
The focus is currently centred on the US, where nearly two and a half years after the Trump administration threatened to ban TikTok in the country unless it divested from its Chinese owners, the Biden administration is now doing the same.
What are the concerns around TikTok?
Explained in two words: data privacy/security.
Recently, there has been growing concern regarding the use of TikTok and its potential impact on data security. Many are worried that the app is collecting sensitive user information and sharing it with the Chinese government. Some worry that the Chinese government could use the app to spread propaganda to a US audience.
The fundamental concern behind both is that any company doing business in China may ultimately be subject to Chinese Communist Party laws, with all its implications.
Lawmakers have renewed their scrutiny of TikTok due to its ties to China through its parent company, ByteDance. This comes after a report last year suggested that US user data had been accessed multiple times by China-based employees. TikTok, however, has disputed this claim, and why wouldn’t they?
What has TikTok said about the accusation?
At a Harvard Business Review conference earlier this month, TikTok CEO Shou Chew reiterated the company’s commitment to address lawmakers’ concerns:
“The Chinese government has actually never asked us for US user data,” Chew said, “and we’ve said this on the record, that even if we where asked for that, we will not provide that.” Chew added that “all US user data is stored, by default, in the Oracle Cloud infrastructure” and “access to that data is completely controlled by US personnel.”
The real question here is, is TikTok truly a threat to national security in the US?
One of the biggest fears is not being able to have full control over the matter.
China’s national security laws compel companies under its jurisdiction to cooperate with various security activities. However, it is difficult for the public to verify the extent of this cooperation.
And although TikTok doesn’t operate in China directly, its parent company, ByteDance does. There’s a potential leverage the Chinese government may have over them.
Overall, privacy and security researchers have found that TikTok’s app is not inherently malicious or spyware, but there are concerns about the relationship between TikTok and ByteDance with the Chinese government.
Source: Deposit Photos
Some exciting updates are coming to the platform for marketers and B2B owners that are regular users of the platform. Aside from the long-awaited native scheduling feature, LinkedIn has released some interesting updates regarding the company pages.
- Native scheduling: You will be able to choose what day and time you want your post to go live when you compose it on the web. With LinkedIn’s native post scheduling option, you can plan your posts up to three months in advance.
- Get the most reach for job listings: Company pages can now share an automated post when they add an open job listing, which helps maximize exposure and reach.
- Audio Events: LinkedIn is introducing audio events for Company Pages, providing another way to engage with your on-platform audience. With this new feature, you can make professional connections via audio events or listen to industry-specific discussions with leaders in your field. This addition could be valuable given the platform’s user base, even if broader interest in audio options has decreased.
- Connecting businesses together: As of now, Company Pages will have the ability to follow other pages in the app, giving you another way to stay informed on relevant industry trends in real-time.
Source: Social Media Today
Meta is advancing its advertising options to help businesses grow even further by leveraging Reels’ growing popularity on Facebook and Instagram. Recently, Meta introduced some new ad video options, allowing advertisers to take advantage of short-form video clips to maximize their promotional opportunities.
Firstly, Meta is introducing Click-to-Messenger ads within Facebook Reels. This will enable businesses to drive traffic directly to DMs via Reels clips. Click-to-Messenger ads in Reels will include a large CTA button to drive contact via private chat. This feature capitalizes on the increasing popularity of short-form video and private message interaction.
Secondly, Meta is making WhatsApp Conversion optimisation available for Facebook Reels Ads. This addition further enhances the capabilities of Reels Ads for businesses.
As mentioned by Meta:
“Typically, we show Click to Messenger ads to people who are most likely to initiate a conversation with a business on WhatsApp, Messenger or Instagram. With this update, advertisers who use the Sales, Engagement or Traffic objectives to add a “Send Message” button to their Facebook Reels ads will now give people an option to start a conversation in WhatsApp right from the ad.”
What’s the possible impact for businesses from these updates?
More social media users are now sharing content in DMs instead of public news feeds, causing a major shift in online interaction. A recent survey by Meta showed that almost 50% of Reels viewers contacted a brand through DMs after seeing its short-form video content.
Meta’s broader vision to maximise its business messaging opportunities is also reflected in adding WhatsApp to the mix. This is an area where Meta still holds a significant lead over its competitors.
Meta is also making the Ads on Facebook Reels post engagement objective more widely available, which will provide another means to measure reactions, comments, and likes, and compare them to engagement from other ads or campaigns.
As a result, the Meta In-Stream Reserve video ad product is being retired from its in-Stream video options due to the growth of short-form video options.
1. New Product Update: new segmentation feature
Source: Pexels | Luis Gomes
We strive to continuously enhance our customer experience, and one way we do so is by constantly improving our marketing automation platform.
Our technical team have recently deployed a new version of database, which enables to segment users also “excluding” them based on any behaviour. This capability has been present for a long time on our triggered campaigns, however we needed to wait this update to have it on any type of campaign.
This not only creates the ability to create specific and impactful campaigns, but also allows our clients to target users who have not yet completed desired actions.This new product update also permits “user merge”, which allows our customers to track users on cross devices.
User merge is particularly useful when considering the increasing number of devices individuals now own, making it more difficult to track users and consolidate data. By merging user data, we can better understand user behaviour on cross-device and provide more personalised experiences.
Additionally, this update allows us to enhance our data analysis capabilities, giving us greater insights into the effectiveness of our campaigns and how to improve them moving forward.
As Frizbit team, we attended the e-Show Barcelona 2023, the largest event for eCommerce and Digital Marketing in Barcelona.
During the conference, we participated in a variety of workshops and seminars, gaining valuable insights into topics such as cookieless strategies for 2023, mobile commerce, hyperpersonalisation, social media marketing, and customer experience optimization
We also had the opportunity to network with other industry professionals, establishing new connections and paving the way for future collaborations.
Overall, our attendance at e-Show Barcelona 2023 was an enriching experience that has provided us with a wealth of knowledge and inspiration!