{
  "id": 70671,
  "title": "Are we even sure visitors will return ?",
  "url": "/competitions/ga-customer-revenue-prediction/discussion/70671",
  "author_name": "Y Assous",
  "post_date": "2018-11-06T13:14:03.046000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello all,</p>\n\n<p>We've seen in thet test data v1 that the number of visitors purchasing is very low. With the new rules of the competition, are we even sure that a visitor will get back on the website AND purchase anything ? </p>\n\n<p>I'm asking that because the new rules seems to change the competition almost as a classifcation problem (like Oliver showed us) than predicting how much a visitor will purchase.. </p>\n\n<p>Let's imagine that only 0,0001% (random number) of visitors  that purchased in our train data v2  will purchase in the future dataset...</p>\n\n<p>Anyone have ideas on how we (competitors) can handle it ? </p>",
  "messages": [
    {
      "id": 416242,
      "postDate": "2018-11-06T13:14:03.047Z",
      "content": "<p>Hello all,</p>\n\n<p>We've seen in thet test data v1 that the number of visitors purchasing is very low. With the new rules of the competition, are we even sure that a visitor will get back on the website AND purchase anything ? </p>\n\n<p>I'm asking that because the new rules seems to change the competition almost as a classifcation problem (like Oliver showed us) than predicting how much a visitor will purchase.. </p>\n\n<p>Let's imagine that only 0,0001% (random number) of visitors  that purchased in our train data v2  will purchase in the future dataset...</p>\n\n<p>Anyone have ideas on how we (competitors) can handle it ? </p>",
      "rawMarkdown": "Hello all,\n\nWe've seen in thet test data v1 that the number of visitors purchasing is very low. With the new rules of the competition, are we even sure that a visitor will get back on the website AND purchase anything ? \n\nI'm asking that because the new rules seems to change the competition almost as a classifcation problem (like Oliver showed us) than predicting how much a visitor will purchase.. \n\nLet's imagine that only 0,0001% (random number) of visitors  that purchased in our train data v2  will purchase in the future dataset...\n\nAnyone have ideas on how we (competitors) can handle it ? "
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "416242": "Hello all,\n\nWe've seen in thet test data v1 that the number of visitors purchasing is very low. With the new rules of the competition, are we even sure that a visitor will get back on the website AND purchase anything ? \n\nI'm asking that because the new rules seems to change the competition almost as a classifcation problem (like Oliver showed us) than predicting how much a visitor will purchase.. \n\nLet's imagine that only 0,0001% (random number) of visitors  that purchased in our train data v2  will purchase in the future dataset...\n\nAnyone have ideas on how we (competitors) can handle it ? "
  }
}