How many Google reviews you need to raise your rating is not a count. It's a rate: how many a month, for how long, and what those reviews average.

Google displays one decimal and rounds at .x50, so a listing showing 4.7 has a true average anywhere between 4.650 and 4.749. Every free Google review calculator treats that displayed number as exact. That's why they tell a restaurant with 145 reviews it needs 146 more, when the real answer is 35.

To display 5.0 on Google you need a true average of at least 4.95. If the reviews you're collecting average below your target threshold, no volume of them ever gets you there, however long you keep going. Your rating drifts toward what you're currently collecting, not toward 5.

At a 4.9 rating, one 1 star review takes 39 five-star reviews to cancel out. At 4.0 it takes three. How many reviews you already have makes no difference to that number.

There's an exact answer, and it isn't one number. It's three: how many reviews you collect a month, how long you keep it up, and what those reviews average.

Drop any one of the three and what's left stops being useful. "You need 146 five-star reviews" doesn't tell you whether that's four months of work or four years, and it assumes every single one of them is perfect.

Start with where you are, because the same tenth of a star costs wildly different amounts of work depending on how many reviews you already have.

Perfect reviews needed to gain 0.1 stars

Read the 4.6 row. On 100 reviews, 34 perfect ones move you a tenth of a star. On 5,000 reviews the same tenth costs 1,667. Your review history is a weight, and it gets heavier every month you collect.

So a useful answer sounds like this: 40 reviews a month, averaging 4.7, takes you from 4.5 to 4.6 in about seven weeks. That's something you can staff and schedule. A raw count isn't.

That's the calculator in Reviewflowz. It takes a target, a monthly volume and the average you expect those reviews to hit, then projects your rating forward and shows you where you land.

Google shows one decimal, and it rounds at .x50. So the number on your listing is a band, not a value.

A listing showing 4.7 has a true average anywhere from 4.650 to 4.749. Two businesses can both display 4.7 and be a year of work apart, because one of them sits at the bottom of the band and the other is a rounding error away from 4.8.

That's why your own rating is the one number you can't read off your own listing. And it compounds, because every plan you make from the displayed number inherits the error.

Two things, and they're separate problems. The numbers are wrong. And where the numbers happen to be right, they still aren't useful.

Here's a test you can run yourself. La Casa des Jumelles is a tapas restaurant in Aix en Provence. It has 145 reviews, Google displays 4.9, and its true average is 4.938. How many perfect reviews does it need to display 5.0?

The answer is 35.

Here's what the free calculators return instead.

Every one of them quotes this restaurant at least four times the work it actually needs, and one of them quotes it thirty seven times over.

The reason is the same in each case, and Dalton Luka shows it most clearly because it's the one that claims to handle rounding. It targets 4.95, correctly, because that's what displaying 5.0 requires. Then it multiplies 4.9 by 145 to get the restaurant's current star total. It applies the rounding rule to the destination and ignores it at the origin. Feed that same calculator the true 4.938 and it returns 35.

The arithmetic engine is fine. The number going into it is wrong.

ReviewTrackers makes the point even faster. Send it 4.9 and it says 146. Send it 4.938 and it says 35. One digit of input, and the plan goes from a year to a month.

Every one of these calculators assumes each new review is a perfect five. Nobody collects only fives. If you've been averaging 4.6, planning around a stream of nothing but 5.0s produces a date you're going to miss.

Almost none of them gives you a date at all. You get a count, with no sense of whether it's a quarter of work or a decade.

And not one of them tells you when your target is out of reach. Ask for a 5.0 and Yotpo divides by zero and prints "∞ 5-star reviews needed to achieve a 5 star rating". ReviewTrackers returns a WordPress error page. Kzoom quietly rewrites your 5.0 to 4.95 and answers as though that's what you asked for.

Not one of them asks how many reviews you collect a month, and not one asks what those reviews are currently averaging. Those are the two numbers that decide the answer.

You can't fix any of this with better arithmetic. To know your true average you need the exact breakdown of your reviews: how many ones, how many twos, how many threes, fours and fives. Without that histogram there's no way to do this properly, and every calculator that asks you for a rounded rating and a total is guessing at the number everything else depends on.

To display 5.0 you need a true average of at least 4.95. The rule generalises: to display any rating, you need a true average of that rating minus 0.05.

With perfect reviews only, here's what 5.0 costs.

A business at a true 4.7 with 500 reviews needs 2,500 consecutive five-star reviews to display 5.0. Not 2,500 reviews. 2,500 perfect ones, with nothing else mixed in.

Which brings us to the part no free calculator will tell you. If the reviews you actually collect average below your target threshold, you never get there. Not slowly. Never.

Your average doesn't drift toward 5. It drifts toward whatever you're currently collecting. Take a business sitting at 4.2 across 200 reviews, collecting 20 reviews a month that average 4.6. After a year it's at 4.418. After five years of that it's at 4.543, and it still hasn't reached 4.6.

So volume isn't the lever you think it is. Quality is.

Better than you probably think, and 5.0 is almost certainly not what you want.

Anderson and Magruder studied this in The Economic Journal in 2012 and found that an extra half star causes restaurants to sell out 19 percentage points, or 49%, more often. The way they proved it is the interesting part for us: they compared restaurants with nearly identical true averages that happened to land on opposite sides of a rounding cutoff. The rounded number is what moves customers, which is exactly why starting your planning from it is the wrong move.

Northwestern's Spiegel Research Center found that purchase likelihood peaks somewhere between 4.0 and 4.7, then falls again as ratings climb toward 5.0. That's e-commerce data rather than local business data, so treat it as a direction rather than a target.

Put those together and the practical answer is to aim for the next band up, not the top of the scale. A 4.8 that's climbing is worth more than a 5.0 you'll spend three years failing to reach.

To get back to the average you had before a 1 star landed, you need this many five-star reviews:

The surprising part is what's missing from that table. Your review count. It isn't there because it cancels out. The answer is the same whether you have 50 reviews or 50,000.

Here's the proof at 4.9. A hundred reviews averaging 4.9 is a star total of 490. The 1 star makes it 491 across 101 reviews, which is 4.8614. Add 39 five-star reviews and you're at 686 across 140, which is 4.9000 exactly. Run the same arithmetic at 1,000 reviews and the answer is still 39.

So the better your rating, the more one bad review costs you. At 4.0 it costs three reviews. At 4.9 it costs thirty nine. That's the trap of a high rating, and it's why collecting your way past a bad review is usually the most expensive option available to you. Getting the customer to change it costs one conversation instead of thirty nine reviews, and getting it removed takes it out of the average altogether, which no amount of new reviews can do.

The number of five-star reviews you need is not something you can act on tomorrow. Here's a number that is.

Take the reviews you collected in the last 30 days and average them. Compare that to your overall rating.

It's the only figure here that reacts to what you did last week. Your overall rating is a decade of history with a month of effort mixed in, so it moves too slowly to steer by. The 30 day average moves the moment your collection does, which makes it the one worth putting on a wall.