How to choose a SaaS plan, and how Shoryaku AI's three plans differ

You open the pricing table, compare the 3 plans lined up there, and end up picking the middle one. I think that is common.

You cannot decide because the table does not say whether an item will matter for you. What the pricing table writes down is whether an item is there or not. It is not because you are indecisive.

The first half of this article is how to read a plan table, usable for any SaaS. The second half explains Shoryaku AI's 3 plans (a marketing-support SaaS, https://shoryaku.net ) with that way of reading.

3 ways to look at a pricing table

1. When there are 2 or more limits, the one that runs out first is the real limit

Next to "analysis, 100 times a month," there is sometimes another limit, such as "credits up to ◯◯." In that case, what you can actually use is the smaller one.

On a service that uses AI, this shape almost always appears. The weight of 1 run differs every time. A business with little data, and a business with hundreds of campaigns in the ad account, consume something completely different for the same "1 run."

So comparing the counts alone is meaningless. What you should look at is how many times you can run if the heavy processing continues. A pricing table that does not write that down may be showing the count side as larger than it is.

2. Separate whether the gap to the higher plan is volume, features, or quality

Sorting the gap of a higher plan into these 3 makes the decision faster.

The one that is easiest to judge wrong is quality. You cannot compare it on the catalog table, so if there is a free period, check the quality before you check whether the volume is enough.

3. Whether you can change mid-way, and what happens when you do

Checking this before you subscribe makes the first choice lighter.

• Can you move up and down at any time? Does it switch immediately, or from the next month?
• Is the difference prorated, or do you pay the whole amount again?
• When you cancel, can you use it until the end of the period, or does it stop on the spot?

If you know you can change it, the conclusion is to start from the lower plan. If it is an annual contract and you cannot change it mid-way, the time you spend on the first choice is worth more.

Shoryaku AI's 3 plans differ in only 3 ways

I apply the reading above to our own plans as it is. What differs is only these 3. All of the amounts are monthly.

Everything else is the same on every plan.

• AI chat
• Generating improvement proposals (no count limit. Within the remaining credits)
• External connections (no limit on the number. Google Analytics, Search Console, Google Ads, Meta Ads, Instagram)
• MCP connection (use it from Claude Code, Codex, Cursor, and Gemini CLI)
• Drafts of blog posts and social posts (in chat, ask it to "write with 〇〇 as the theme")
• Generating and printing the monthly review
• The check against the acquisition-cost ceiling, and the confidence calculation for proposals (explained in another article)

The AI model is not changed between Lite and Standard. If the cheapest plan is the one that shows the worst output, the evaluation of the service is fixed there for the person who chose that plan. What is limited is only the count. The quality of what comes out is the same.

"20 times a month" is not a guarantee

I apply way of looking 1 to our own product as well. Shoryaku AI also has 2 limits.

Separate from the analysis count, there is a limit on AI usage (credits). If heavy analyses continue, the credits run out before the count does. The more channels you have connected, and the larger the ad account, the heavier 1 run is.

The credit limit is a figure set by placing 25% of each plan's revenue as the ceiling on AI cost (the calculation is written in the cost article). This ratio has not been moved even when the price was cut.

Pro's analysis count is 25 for the same calculation. A model with higher reasoning performance costs about 4 times as much per run. The typical analyses that can be run within the cost ceiling of the ¥4,980 monthly price are 29, so it is stopped at 25 so that the displayed count is one you can actually reach. You could write "100 times a month," but that does not hold at ¥4,980.

The choice comes down to 3 questions

Question 1: Are you running search ads?

If you are not, there is almost no reason to choose Pro.

Pro's gap is the detail of ad analysis, and the model. If there is no ad data, the difference in detail cannot appear.

Question 2: In ads, do you work below the campaign level?

This is the fork.

If you are judging at the whole account and at campaign level, Standard is enough. That is the use where you look at budget allocation and pause campaigns whose results are poor.

If you look at and fix ad groups, ad text, keywords, negative keywords, and creatives, that is Pro. On Lite and Standard, data at this level is not passed to the AI. The screen says "Showing the basic analysis of the account and campaigns," and anything below that is outside the analysis.

The feature for finding lower-competition keywords is also Pro only.

Question 3: How many times a month do you rerun the analysis?

• A pace of once a week is 4 to 5 times a month. Lite is enough
• If you want to look at each channel separately, or review every time you run a move, that is Standard
• Pro's 25 is a number placed as a result of raising the quality, more than for the sake of the count

If you are unsure, start with Lite

You can change the plan at any time.

• When you move up, it switches immediately, and the difference is billed prorated on the spot
• When you move down, it also switches immediately, and the amount overpaid becomes a balance
• Cancellation can be used until the end of the period. There is no prorated refund

So you do not need to pick Pro first and see how it goes. If you use it and the count is not enough, move up. If you want to work on the finer levels of the ads, move up. That is enough.

What to check before you subscribe

• There is no free plan. The only thing you can use for free is a 14-day trial (the reason is here)
• Pro's gap is not that features are added. It is that the grain you can see gets finer. If the business is not running ads, I think the difference is hard to feel
• The analysis count is a ceiling, not a guarantee. If heavy analyses continue, the credits run out first
• There is no prorated refund when you cancel. You can use it until the end of the period

To try it

The 14-day free trial can be tried with no card registration at the start. Think of it as the period for checking the quality above.

If you want to look at the numbers first, there is also a CSV sales diagnosis that does not require signup.

• Product
• Pricing
• SaaS