I've always wondered if a grinding strategy was valid for writers: quality over quantity or quantity over quality, that's the question. Quantity is a straightforward metric: the number of posts written on a blog, but how do you measure quality?

I would say that quality is inversely proportional to the amount of time that takes a user to satisfy a query:

$Q_{article} = \frac{1}{\Delta {t}}$

For example, if you search 4 + 4 at Google, it returns 8 in a few ms. Google doesn't try to explain where maths are coming from; it solves the query as quickly as possible. Furthermore, if you type the expression on the Chrome navigation bar, you don't even need to press enter: $\Delta {t} \approx 0$.

The article must also provide correct and provable information. If the information is returned quickly but is false, the quality of the article is null.

Assuming that you try to optimize the quality of your articles, is the number of posts correlated to the number of readers?

Let's see some numbers of this site.

If I place the data into a single JavaScript object:

javascript`{ "2019-07": {sessions: 1291, posts: 4}, "2019-08": {sessions: 1052, posts: 5}, "2019-09": {sessions: 1174, posts: 7}, "2019-10": {sessions: 2072, posts: 22}, "2019-11": {sessions: 1708, posts: 30}, "2019-12": {sessions: 3736, posts: 37}, "2020-01": {sessions: 1965, posts: 42}, "2020-02": {sessions: 1464, posts: 59}, "2020-03": {sessions: 1134, posts: 67}, "2020-04": {sessions: 1691, posts: 77}, "2020-05": {sessions: 2434, posts: 90}, "2020-06": {sessions: 3076, posts: 98}, "2020-07": {sessions: 3929, posts: 121}, "2020-08": {sessions: 4401, posts: 146}, "2020-09": {sessions: 5715, posts: 177}, "2020-10": {sessions: 6813, posts: 207}, "2020-11": {sessions: 6899, posts: 238}, "2020-12": {sessions: 4684, posts: 266}, "2021-01": {sessions: 4928, posts: 278}, "2021-02": {sessions: 4993, posts: 284}, "2021-03": {sessions: 5022, posts: 299}, "2021-04": {sessions: 4410, posts: 321}, "2021-05": {sessions: 6138, posts: 342}, "2021-06": {sessions: 6716, posts: 369}, "2021-07": {sessions: 7105, posts: 382}, "2021-08": {sessions: 8010, posts: 382}, "2021-09": {sessions: 7858, posts: 393}, "2021-10": {sessions: 9565, posts: 415}, "2021-11": {sessions: 8675, posts: 434} }`

Now I want to know the relation between the sessions and the number of posts. So, for every month, I calculate a coefficient:

${c_{sp_{i}}} = \frac{sessions_i}{posts_i}$

javascript`{ "2019-07": { "csp": 322.75}, "2019-08": { "csp": 210.4}, "2019-09": { "csp": 167.71428571428572}, "2019-10": { "csp": 94.18181818181819}, "2019-11": { "csp": 56.93333333333333}, "2019-12": { "csp": 100.97297297297297}, "2020-01": { "csp": 46.785714285714285}, "2020-02": { "csp": 24.8135593220339}, "2020-03": { "csp": 16.925373134328357}, "2020-04": { "csp": 21.961038961038962}, "2020-05": { "csp": 27.044444444444444}, "2020-06": { "csp": 31.387755102040817}, "2020-07": { "csp": 32.47107438016529}, "2020-08": { "csp": 30.143835616438356}, "2020-09": { "csp": 32.28813559322034}, "2020-10": { "csp": 32.91304347826087}, "2020-11": { "csp": 28.98739495798319}, "2020-12": { "csp": 17.60902255639098}, "2021-01": { "csp": 17.72661870503597}, "2021-02": { "csp": 17.580985915492956}, "2021-03": { "csp": 16.795986622073578}, "2021-04": { "csp": 13.738317757009346}, "2021-05": { "csp": 17.94736842105263}, "2021-06": { "csp": 18.200542005420054}, "2021-07": { "csp": 18.599476439790575}, "2021-08": { "csp": 20.968586387434556}, "2021-09": { "csp": 19.994910941475826}, "2021-10": { "csp": 23.048192771084338}, "2021-11": { "csp": 19.98847926267281 } }`

I calculate the expected value:

${E(c_{sp_{i}})} = \frac{1}{M_{months}} \sum{c_{sp_{i}}} = \frac{322.75 + 210.4 + \cdots + 19.98847926267281}{29} \approx 51.75$

The result is biased by the first months, where the coefficient was high due to the low number of posts.

Let's take the worst-case scenario, ${c_{sp_{min}} \approx 13.74}$

In the worst-case scenario, after writing 1000 posts, $1000 \cdot c_{sp_{min}} \approx 13740$ sessions per month are expected.

Taking the average coefficient of the last 12 months: ${E(c_{sp_{12 months})} \approx 18.52}$

In a more realistic scenario, after writing 1000 posts, $1000 \cdot c_{sp_{12 months}} \approx 18520$ sessions per month are expected.

If eventually the blog reaches that amount of posts, we'll check the predictions accuracy.

Hi, I'm Erik, an engineer from Barcelona. If you like the post or have any comments, say hi.