By Molly Ploe
on 14-11-2023Social media is a unique marketing vehicle for a number of reasons. What other communication channel can be used to joke around with customers, swiftly solve customer service issues or joke about a marketing campaign where you killed off your mascot to be reborn as a baby peanut?
Social media is the wild west of digital marketing opportunities.
Social media is also unique in how rapidly it evolves.
When Twitter was founded in 2006, people were instantly enamored with the ability to quickly communicate with others. Within a decade, the speed of responses wasn’t a novelty – it was an expectation. According to the 2022 Sprout Social Index™, 13% of U.S. consumers expect brands to respond within the first hour of reaching out on social media, while the other 76% expect a response in the first 24 hours.
Social media continues to rapidly change today. New platforms are introduced, and others are consolidated. Once-beloved platforms can fall from grace (RIP Vine), and new features are introduced all the time to keep users engaged and entertained.
To the social media marketer, the fast-paced nature of these platforms is equal parts exciting and challenging. Staying relevant on social media is a must, and to do that, it’s critical to be tuned in to changes as they happen.
Reading industry blogs is one way to stay at the forefront of change. Here are some of our favorites:
10-11-2023
I don’t know about you, but I find that I can’t turn my business analysis brain off. I find myself wanting to improve just about every process that I experience, and I often find myself conducting business analysis on the processes that I experience as a customer. This happened to me recently when a company asked me for a physical ‘wet signature’ on a form. This wouldn’t be unusual if I was in the same physical location as the person requesting the signature, but I wasn’t—they were literally going to email me a PDF form for me to print out, sign, scan in and email back. Even though Adobe PDF has a signature function (and I have a graphics tablet, so I can literally do the same signature electronically), this wasn’t good enough. It had to be old-school pen on paper. I eventually complied, deciding that I’d rationalize it by calling the process “retro” and “vintage”… Badly “Improved” Processes Might Be MORE Risky: In situations like this, I always want to ask ‘why’ to understand the underlying reasons that things work in a particular way. In this case, if we were to ask why I suspect the underlying answer would be that an old paper process had been replaced, with each step being recreated electronically. In this context, ultimately, a signature acts as a way of authorizing something to happen. You can almost imagine the conversation with a group of Subject Matter Experts (SMEs). One of them is adamant that we couldn’t possibly accept electronic signatures. The legal & compliance SME says that electronic signatures are completely valid, but there is reluctance from other stakeholders for other reasons. Perhaps there’s a perception that by getting a physical signature there is less risk… or perhaps there are other reservations. Some might be genuine, others might be founded on unsubstantiated fears or assumptions. Left unchecked, there is a risk that ‘we’ll do things the way we’ve always done them, just in an electronic format’. This can lead to the worst of both worlds where risk is increased and customers are inconvenienced: Alternation: I could have altered the PDF before I printed it and signed it (it’s relatively easy to edit a PDF). Unless they check it word-for-word they’ll never know, and since it’s scanned, auto-comparing will be difficult. Verification: They didn’t actually have my signature on file. If they weren’t comparing it against anything, then really what is the point? I could have scrawled any old signature and it probably would have been accepted. Security: Unencrypted email isn’t a secure transmission format. Even though it was relatively low-risk mundane information, it’s liable to interception en route. Plus emails can be spoofed so there’s a risk of a customer being impersonated. Scanners: How many customers have scanners? What about people who just take photos on their phone, will that be sufficient? These are just four examples, but they illustrate a key point: The process probably isn’t achieving what it actually set out to do. Yes, you are getting a physical signature. But if the aim is to get a secure, authenticated nonreputable authorisation for something… then the process is failing! Always Ask Why: “Good” Questions Make A Difference The key to avoiding situations like this is to ask why, in varied ways, lots of times. Do this and you’ll get to the core purpose of a process, or process step. In their book “Mastering the Requirements Process: Getting The Requirements Right”, James & Suzanne Robertson call this “The Essence”. In this example ‘getting physical signature’ might be the current step, but the essence is ‘authorize transaction’ (or whatever).A key point here is that if you understand the essence, you can question any underlying assumptions or business rules. It’s possible to ask “how else might we be able to do this”. If the aim is to ‘authorize transaction’ then there are countless other ways of doing this that are more secure and verifiable than a scanned PDF in an email. You could even use the Brown Cow model to question any underlying assumptions that have been made. Asking these questions will help encourage stakeholders to think about the true essence of the process, and about how it might change in the future. A half hour discussion now might save tens of thousands of processing time later, once the process is implemented. This is yet another area where BAs add significant value by helping to ensure things improve in a way that maximizes the benefits that will be delivered both to the organization, and to its customers.
14-11-2023
Hi! Previously we have discussed chatbots. Briefly, chatbots can be categorized into 2 branches: - Retrieval based - Generative Retrieval chat bots rely on a database to search in. However, generative chat bots rely on a model to generate its answers. Generative chat bots require a huge amount of data to be trained, they also require huge resources in order to train them. That makes generative chat bots harder to find and create. Also, their outputs are not perfect, they sometimes make no sense and they make the silliest grammatical mistakes. In this tutorial, we are going to train a generative chatbot using a small dataset. we shouldn’t really expect much from that bot, but the same process and model can be used to train bigger and better models. For this task, we are going to use the Frames dataset from Maluuba. This dataset contains almost 1400 conversations between humans that are trying to reserve travel packages. The data set isn’t designed to be used for training chatbots, however, it can be used to extract simple one-liner conversations that we can use to train a chatbot with. I want to focus on the practical part of this task, so I’m going to leave the theoretical part to articles that better explain it. I’m not going to explain the sequence to sequence algorithm because there are many resources that explain it, but there aren’t many practical examples, especially those that use the dataset API. You can read Tensorflow’s introduction to machine translation tutorial for an understanding of seq2seq.
