[AWS construction and operation and maintenance] SaaS service SNS campaign tool | Pickles

This time, we spoke with Tanaka and Yamashita of Pickles Inc. about the background to the introduction of Beyond's cloud construction and operation and maintenance services, and the impressions of actually using the service.

--Please tell us about your company's business.

Yamashita : Our company offers two marketing SaaS: the social media campaign tool "Can Tsuku" and the diagnostic content creation tool "Diagnosis Cloud Yomitru".

Tanaka-sama : Can Tsuku is a social media campaign tool whose strength is that it can be used by a single marketer. You can easily run gift campaigns such as "Instant Winning" on social media such as X (formerly Twitter), Instagram, LINE, and TikTok. Yomitru is a tool that allows you to create a variety of diagnostic content. It has been introduced by a variety of companies for product promotions and conversions on e-commerce sites. Its advantage is that anyone can easily create diagnostic content using AI.

Tanaka Minoru, CEO of Pickles Inc., and Yamashita Mio, Manager of Product Development Department

――Please tell us the background that led you to introduce our service.

Tanaka : There was a problem with the slow response speed of the company I had previously asked to do so, so I changed to your company.

Yamashita : Actually, personally, I've been dating for about seven years since I built the site, and because we've been together for a long time, I've gotten used to it, and honestly there were some parts that I thought might be a bit slow.

Tanaka : However, from my perspective, I felt that the response speed was quite slow. For example, even if there was a server-side problem, it could take more than half a day for a response to arrive. It was common practice to take about a month or two from the initial consultation to actually start moving forward, including understanding the situation, calculating costs, suggesting schedules, and specific responses.

Yamashita-sama : I've always been used to that method, so I thought it could be no use. Still, Tanaka said, "Isn't this a major business risk?" and once again I felt that "as a service provider, this slowness is not good..."

Tanaka : I think that server-related issues in particular are a major problem from the customer's perspective. However, it is a high risk that it takes too long to respond, so I asked Yamashita, "Is this really okay as it goes on like this?" and "It could be even faster to maintain and operate other companies."

Mr. Yamashita : That's why, so I looked into the actual situation of other companies that build and maintain servers. When I first started researching, I realized that there were companies that had an overwhelmingly fast response support system. For seven years, I thought it was only natural that "slow response" was not the case, but I felt that it wasn't actually the case.

Tanaka : In the end, we decided to consider switching at this time. The OS and PHP version upgrades also overlapped, and slow responses were even more hindered... If we are to prioritize stable service provision as a company, we have concluded that "we should change to an operating company that will move quickly, based on communication."

--How did you find out about Beyond?

Yamashita : I found out by searching the web. I was talking with Tanaka about what to do with the management company, and after doing some research, Beyond's name first came up. It seems to be one of several companies that I've picked up.

Tanaka : I was aware that I was one of the companies Yamashita made a list.

--In the end, what made you choose our company?

Tanaka : What we valued was "can we respond quickly?" That was a major problem at previous companies. Next was the key point: "How extensive a range of consultations can you please." In this regard, other companies that were listed as candidates responded in a manual manner, saying, "We can handle any consultations up to this level." In comparison, I got the impression that Beyond was flexible in answering small insightful consultations , which was a big deciding factor.

Mr. Yamashita : We don't have any specialized infrastructure engineers, so we have many opportunities to consult about server configuration and operation and maintenance. In such cases, the deciding factor was whether they could easily consult with us or whether they could act quickly when there was a risk. In fact, I spoke to them many times in the meetings in advance, and they had solid technical knowledge and responded quickly. that geekiness led to a sense of security and helped me make the final decision.

Tanaka : In fact, even if we were to ask someone else that would be rejected, he flexibly thought about it, saying, "Let's think about it." That was what impressed me, and I ended up thinking, "This company would likely be able to help with new initiatives."

Yamashita : And, cost was also an important factor. It was also important that we could keep the construction and maintenance costs within the costs set by us, but Beyond was able to make suggestions while keeping them within the expected circumstances.

SNS campaign tool "Can Tsuku"

--What kind of services do you use Beyond's services for?

Yamashita : We ask for the construction and operation and maintenance of "Cancel". Can Tsuku is a system that collects posts with hashtags and reposted posts, and is characterized by its compatibility with a large number of social media.

Tanaka : Specifically, there are four types: X, Instagram, LINE, and TikTok. Since both data are acquired via APIs, when running an instant win-style campaign, you must immediately process and output "wins and misses" on the server. Therefore, if access is concentrated, the server load will be extremely high. Until now, there were no load balancers, so responses could be slow when accessing large amounts of data. Therefore, in this server construction, we have decided to set up a configuration of multiple units with load balancers installed and update the system to a system that can also support autoscaling.

Mr. Yamashita : Regarding some of the compositions, Mr. Beyond suggested, "We have no track record so we will come up with alternatives," but in the end we were able to implement them in the way we wanted. Normally, I think there would be many cases where people would be rejected saying, "If they don't have a track record, they would be tough," but they tried out new methods and provided solid support, which has made me even more confident.

Tanaka : This time, we have optimized the configuration in a way that is in line with our requests, but even if the social media campaign receives a greater response than expected, our servers can withstand. I am truly grateful that I was able to create a solid foundation that can adequately accommodate load-bearing systems like Instant Win.

Pickle's "Cance" environment structure diagram

--What are the specific aspects of Beyond that are useful?

Yamashita : When migrating from an old server to a new server, this was actually the first time I was dealing with PHP's FPM environment, and I needed to make more detailed changes than I thought. Specifically, it includes what to do with the error log output destination and adjusting the PHP middleware. I started to wonder one after another, "What should I do in this situation?", but when I consulted Beyond about this, he told me in great detail, "Well, let's do that here."

Tanaka : The work we requested will also be reported a few hours later, saying that we have already responded. Until now, we often stopped exchanging things like "I'll check it out before contacting you," but we received results right away, so our work was not interrupted.

Yamashita : In fact, I feel that it was really helpful that the work was smooth even during the transition, barely stopping work.

Tanaka : Another helpful thing is the usual surveillance system. There are quite a few situations where the database load is high when you cancel, but it's quick to receive a message in Slack saying, "The load is rising, so we're checking." You will also receive specific information such as "It seems that this SQL query is the cause," so we can respond quickly.

Yamashita : In my previous environment, it was only when a disability occurred that I realized, "Is something happening?" With Beyond, they set the threshold properly and identify bottlenecks before they become a problem. Reports are also made frequently in Slack, so it's easy to think of countermeasures at that time too.

Tanaka : It was also very helpful when our company made a mistake in setting up a cron job and suddenly the load went up. At that time, I received many alerts saying, "The load is incredible! (In fact, it's polite language)," so I was able to quickly realize the problem. If it hadn't been there, I think I would have left it alone for a long time without realizing it.

Yamashita : This has led me to reassess the design itself of the application, so it was a big factor that I was able to not only discover and stop mistakes and achieve fundamental improvements.

Tanaka : We contact you frequently, and not only monitor you, but also analyze that "this is what's causing it." The server load has also been reduced, resulting in stability of the overall service . We feel that it has helped us in many ways.

Case studies of can-take

--Please tell us your impressions after using our service.

Yamashita : The first thing I'm happy about is that I can communicate via Slack rather than via email. Since most of the internal communications are available, it's very helpful to be able to immediately consult and contact us as an extension.

Tanaka-sama : It's reassuring to see that it responds in real time without having to go through the trouble of checking your email. I once again realize that speed is important after all. get the impression that Beyond's "instant response" culture is thoroughly implemented throughout the company

Yamashita : Previously, we were in a situation where we couldn't get any new suggestions or information unless we asked questions. We also have little expertise, so we often don't know what questions to ask. Now, even if we briefly explain what we are struggling, we are getting various suggestions such as, "Well, there is a way to do this." The people in charge share their knowledge with each other, so I feel that they can act so flexibly

Tanaka : I would appreciate it if you could give us advice based on suggestions such as, "This will make it more efficient" and "This setting may be the reason why the load is so high." This can also be improved while learning, "Is that so?"

Yamashita : Also, it's helpful when dealing with the late night hours. In many cases, you have to update late at night to cancel, but with previous companies, there was an additional charge each time, like "Late-night service is an option." The cost was a little high, so it was difficult to make internal adjustments, such as "Would you like to stop it in the middle of the night on another day?"

Tanaka : Nowadays, with Beyond's monitoring and maintenance services, you can request late-night services on weekdays relatively smoothly, so you can take the plunge and complete the process in the middle of the night. As a result, the risk of service outages has been significantly reduced as users can update and maintain the service at a low time.

Yamashita : I'm grateful for that. In fact, there are some parts that we don't know when the time will be when the massive traffic will arrive, but as we communicate, we can immediately say, "Well, let's take a look at it in the middle of the night." This is a speed of response that was a bit unthinkable before, and this is a great relief for service operators.

Diagnostic

--What is expected of us in the future?

Yamashita : First of all, we want to develop "Cance" and "Yomitoru" together with them. We are steadily developing our own development, but we believe that AWS services and infrastructure solutions are still deep. I think it's encouraging to be able to discuss this together and be accompanied by a "runner" so to speak.

Tanaka : I've heard that Beyond has a lot of track record of supporting the game field. I think it's common for gaming companies to concentrate access all at once. If you have the knowledge that has dealt with such cases, I believe you will know how to deal with it even when you run Can and Yomitru grow rapidly. For example, if you say, "We've become 10 times more users of the service," we hope that we'll get a suggestion like, "If we scale this way, we can get through it."

Yamashita : In fact, I am conscious of taking measures to deal with the burden at this stage, but as it actually grows, problems may arise in unexpected areas. It would be encouraging if you could push yourself back at that time, like, "This architecture is great" or "This AWS feature will help you get through it."

Tanaka : We also feel pressured to "grow that stage," but on the other hand, that is also motivated. If Beyond supports you by saying that you can grow as much as you want, you'll be motivated to "do your best and increase your users by 10 times."

Mr. Yamashita : As we continue to expand our services, I think there will be more situations where we will be discussing various kinds of consultations and thinking about implementation together. It would be helpful if you could take the stance of "please talk about anything" in response to this.

Tanaka : Looking at it, I feel that he is an attitude of "We can handle a wide range of issues around AWS," so I am hoping for such a companion.

--Please tell us about your future business development.

Yamashita : My goal is to have more people use Camping and Yomitor. Our company has a policy of "producing quality communication between companies, companies and consumers." In this regard, we would like to continue to hone our functions so that companies adopt it can achieve results that they can feel that they are glad that they requested it here.

Tanaka : Recently, I've updated it to allow users to be managed with Yomitor. For enterprise customers, user management functions are an important factor in the introduction of SaaS, so they meet those needs. We will continue to work on various efforts to "lower the threshold for adoption" and create a service that is easy for any company to "try using it."

Mr. Yamashita : I believe that it is important to "create an environment that will ensure that all customers can use the service with peace of mind." We would like to continue to raise the service level, including the load countermeasures mentioned earlier, security and ease of data management.

Tanaka: Considering that, Beyond's support is encouraging. I would be happy if I could continue to consult with you on various aspects in order to expand my services.

--Thank you very much for your valuable talk today!