Linkedin competitor scraping workflow last node not working correctly

[jimvlieger]

Hello n8n Community,
I’m seeking some assistance with a workflow I’ve built. The goal of the workflow is to scrape LinkedIn posts from a list of creators and save the relevant content into an Airtable base.

The entire workflow seems to be running correctly, but the final Airtable “Create a record” node is not saving any data. It appears that the records are not being created, or if they are, they are empty.

I suspect the issue might be with the configuration of the last node or an earlier data handling step. I have attached screenshots of the workflow and the Airtable configuration for reference.

Could someone please review the workflow and suggest what might be wrong? Specifically, I’m trying to understand why the data isn’t being passed to the final Airtable node correctly.
Thank you in advance for your help!

Please share your workflow

https://drive.google.com/file/d/1qIlQUT-QyYGbxkoa3bYQe20OT63rnFwQ/view?usp=sharing 

[jimvlieger]

@Wouter_Nigrini Hi Wouter, any good suggestions? I shared the workflow.

Jim

[Wouter_Nigrini]

@jimvlieger apologies Ive been offline and busy with work the last few days. Let me go through your question and come back with a proper response. Give me a few hours to respond

[jimvlieger]

Hi @Wouter_Nigrini

Thanks for replying.

The competitor scraping workflow is running but not functioning correctly. When I check the results in Airtable, the output does not sync properly. I suspect that my Apify and Airtable settings are configured incorrectly .. and probably more…

Workflow first part competitor scraping

Workflow second part competitor scraping

Airtable settings - wrong data sync from n8n

Create a record node ( values are probably incorrect )

Apify value is incorrect

Get exiscting post (settings are probably incorrect)

Set source 1 is incorrect?

Set source 2 (Do i miss values?)

Merge 1

Filter out

If new post excist

Switch content

In Apify the actor runs but always fail

I have my own linkedin profile as a test for scraping content but I dont see my content in airtable..

If you could help me sort out this workflow beast, I would really appreciate it

[Wouter_Nigrini]

Can you show me the Merge 2 nodes data in the left panel for this screen

[jimvlieger]

Hi Wouter,

See below

[Wouter_Nigrini]

So most values are null from the Merge which means this is where the issue is from being blank in your excel sheet. what are the variables looking like where you input into the merge? For example where is the likesCount variable being set?

[jimvlieger]

Hi @Wouter_Nigrini,

I’ve checked, and I don’t see the likesCount variable anywhere in the Apify node’s output. I think this data should come from the LinkedIn scraper, right? The Apify node should be the source of the likesCount and other post-related data. What do you think?

[Wouter_Nigrini]

Let me first take a step back and ask if my assumption is correct. Your question was around the excel columns which didnt populate and likesCount was one of them?

If so then yes, since you’re using the Merge Node, you’re merge that value from somewhere. Is this a variable youre looking for to populate in the excel?

Im not sure where likesCount is supposed to come from as I havnt worked with the linkedin node before. Can you maybe share your workflow in a codeblock? My view currently is very limited.

[jimvlieger]

Hi Wouter,

See below link for all necesarry information and code block

LinkedIn AI System

and https://www.youtube.com/watch?v=GmlQZ8MG7HY

The likescount are from the scraped linkedin post so the connection must be from apify into airtable

[Wouter_Nigrini]

So if I understand correctly, the linkedin data is scraped and added into an airtable? Does your airtable have the likesCount value populated?

[jimvlieger]

Yes, its in my airtable

[jimvlieger]

in his toturial it works so I think some nodes need to be re configurate.

i am a little bit lost … Tried already 100 options but its stil not working. Hopefully you can make this automation to work.

[Wouter_Nigrini]

You need to debug the first scraping workflow you have. The values are not being pulled into your airtable.

[jimvlieger]

[
  {
    "urn": "7359117118409191425",
    "full_urn": "urn:li:activity:7359117118409191425",
    "posted_at": {
      "date": "2025-08-07 09:04:06",
      "relative": "1 week ago • Visible to anyone on or off LinkedIn",
      "timestamp": 1754550246813
    },
    "text": "𝗪𝗮𝗮𝗿𝗼𝗺 𝟳𝟯% 𝘃𝗮𝗻 𝗱𝗲 𝗡𝗲𝗱𝗲𝗿𝗹𝗮𝗻𝗱𝘀𝗲 𝗯𝗲𝗱𝗿𝗶𝗷𝘃𝗲𝗻 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀 𝘃𝗲𝗿𝗸𝗲𝗲𝗿𝗱 𝗶𝗻𝘇𝗲𝘁 (en hoe het beter kan)\n\nDe meeste organisaties zetten één AI-assistent in voor alles.. een fundamentele denkfout die miljoenen euro's aan potentiële waarde vernietigt.\n\nWat als je in plaats daarvan een team samenstelt van gespecialiseerde AI agents die naadloos samenwerken?\n\n𝗪𝗲𝗹𝗸𝗼𝗺 𝗯𝗶𝗷 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁 𝗼𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗶𝗲: de strategie die toonaangevende Nederlandse bedrijven implementeren om hun concurrenten te overtreffen.\n\nHet principe is eenvoudig maar krachtig:\n\n• 𝗧𝗿𝗮𝗱𝗶𝘁𝗶𝗼𝗻𝗲𝗹𝗲 𝗮𝗮𝗻𝗽𝗮𝗸: Eén algemeen AI systeem dat worstelt met complexe processen\n\n• 𝗢𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗶𝗲 𝗮𝗮𝗻𝗽𝗮𝗸: Meerdere gespecialiseerde AI agents die samenwerken onder centrale coördinatie\n\nEen Nederlandse logistieke speler implementeerde dit jaar een orchestratie laag die 5 verschillende AI-agents coördineert:\n- 𝗩𝗼𝗼𝗿𝗿𝗮𝗮𝗱𝘃𝗼𝗼𝗿𝘀𝗽𝗲𝗹𝗹𝗶𝗻𝗴𝘀 𝗮𝗴𝗲𝗻𝘁\n- 𝗥𝗼𝘂𝘁𝗲 𝗼𝗽𝘁𝗶𝗺𝗮𝗹𝗶𝘀𝗮𝘁𝗶𝗲 𝗮𝗴𝗲𝗻𝘁\n- 𝗗𝗼𝗰𝘂𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗲 𝗮𝗴𝗲𝗻𝘁\n- 𝗞𝗹𝗮𝗻𝘁𝗲𝗻𝘀𝗲𝗿𝘃𝗶𝗰𝗲 𝗮𝗴𝗲𝗻𝘁\n- 𝗙𝗮𝗰𝘁𝘂𝗿𝗲𝗿𝗶𝗻𝗴𝘀 𝗮𝗴𝗲𝗻𝘁\n\nResultaat? 43% snellere doorlooptijden en 35% kostenbesparing binnen zes maanden.\n\n𝗗𝗲 𝗱𝗿𝗶𝗲 𝗸𝗿𝗶𝘁𝗶𝗲𝗸𝗲 𝗯𝗼𝘂𝘄𝘀𝘁𝗲𝗻𝗲𝗻 𝘃𝗼𝗼𝗿 𝘀𝘂𝗰𝗰𝗲𝘀𝘃𝗼𝗹𝗹𝗲 𝗼𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗶𝗲:\n\n1. Taaksplitsing: Verdeel complexe werkstromen in duidelijk afgebakende deeltaken\n\n2. Specialisatie: Train iedere agent om uitmuntend te zijn in één specifiek deelproces\n\n3. Communicatieprotocol: Ontwikkel een gestandaardiseerde manier waarop agents informatie uitwisselen\n\nBedrijven die deze aanpak omarmen, creëren een exponentieel schaalbare automatisering laag die menselijke teams bevrijdt voor strategisch werk.\n\nWelke complexe bedrijfsprocessen zou je transformeren met een georchestreerd AI team?\n\n#ai #aiagent #automation #openai #workflow #emailautomation #aiagents #techstack #efficiency #leadgen #B2BMarketing #LeadGeneration #MarketingAutomation #EmailMarketing #DigitalMarketing #GrowthHacking #MarketingStrategy #MachineLearning #n8n #GenerativeAI #google #hr #sales #cms #dxp #composable #chatgpt #marketing #hrm #finance #klantenservice",
    "url": "https://www.linkedin.com/posts/jimdejager_ai-aiagent-automation-activity-7359117118409191425-l42g?utm_source=social_share_send&utm_medium=member_desktop_web&rcm=ACoAAF3IbmsBapYWOUSHvOaplfrho6JFb1hIMXM",
    "post_type": "regular",
    "author": {
      "first_name": "Jim",
      "last_name": "de Jager",
      "headline": "Building AI Bots & Automations 🏆➤ Consultative Sales | Solution Sales | Presales | Customer Success & Change Management",
      "username": "jimdejager",
      "profile_url": "https://www.linkedin.com/in/jimdejager?miniProfileUrn=urn%3Ali%3Afsd_profile%3AACoAAAwYQP0BOafQGTJtNJDighWarr4fn-EvQkA",
      "profile_picture": "https://media.licdn.com/dms/image/v2/D4E35AQEkQbsQSGOprw/profile-framedphoto-shrink_800_800/B4EZflg2q5GwAk-/0/1751902284898?e=1756033200&v=beta&t=69CPt5S_6Vf7C6pR3ybUSy14_76FH8La3YIS9kWS1vo"
    },
    "stats": {
      "total_reactions": 11,
      "like": 11,
      "support": 0,
      "love": 0,
      "insight": 0,
      "celebrate": 0,
      "funny": 0,
      "comments": 0,
      "reposts": 0
    },
    "media": {
      "type": "image",
      "url": "https://media.licdn.com/dms/image/v2/D4D22AQFRGf6sn7Wq0A/feedshare-shrink_2048_1536/B4DZiDWBBMGkAo-/0/1754550245183?e=1758153600&v=beta&t=zeOQoGg0gzUIMm52ucWY3b4ZbLGTVeTYc6unzeaE2gM",
      "images": [
        {
          "url": "https://media.licdn.com/dms/image/v2/D4D22AQFRGf6sn7Wq0A/feedshare-shrink_2048_1536/B4DZiDWBBMGkAo-/0/1754550245183?e=1758153600&v=beta&t=zeOQoGg0gzUIMm52ucWY3b4ZbLGTVeTYc6unzeaE2gM",
          "width": 1024,
          "height": 1024
        }
      ]
    },
    "profile_input": "https://www.linkedin.com/in/jimdejager/"
  },
  {
    "urn": "7358063647660621824",
    "full_urn": "urn:li:ugcPost:7358063646935027712",
    "posted_at": {
      "date": "2025-08-04 11:17:59",
      "relative": "1 week ago • Edited • Visible to anyone on or off LinkedIn",
      "timestamp": 1754299079814
    },
    "text": "𝗛𝗼𝘄 𝗜 𝗕𝘂𝗶𝗹𝘁 𝗮𝗻 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁 𝗧𝗵𝗮𝘁 𝗪𝗿𝗶𝘁𝗲𝘀 𝗮𝗻𝗱 𝗣𝗼𝘀𝘁𝘀 𝗩𝗶𝗿𝗮𝗹 𝗟𝗶𝗻𝗸𝗲𝗱𝗜𝗻 𝗖𝗼𝗻𝘁𝗲𝗻𝘁 (full demo!)\n\nMost companies struggle with one recurring challenge: staying visible on LinkedIn without spending hours every day creating content.\n\nMy solution? An AI automation that handles everything from idea to post:\nIt generates new post ideas based on trends and my previous content.\n\nIt designs on-brand visuals automatically.\n\nIt schedules and publishes posts on LinkedIn completely hands off.\n\nThe result: a consistent stream of engaging content without the manual grind of writing and designing.\n\nThe essence? The future of personal branding isn’t about working harder. It’s about automating smarter.\n\n𝗥𝗲𝗮𝗱 𝘁𝗵𝗲 𝗳𝘂𝗹𝗹 𝗮𝗿𝘁𝗶𝗰𝗹𝗲 𝘁𝗼 𝗴𝗲𝘁 𝗱𝗲𝗲𝗽𝗲𝗿 𝗶𝗻𝘀𝗶𝗴𝗵𝘁𝘀 𝗶𝗻𝘁𝗼 𝘁𝗵𝗲 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄.\n𝗪𝗮𝘁𝗰𝗵 𝘁𝗵𝗲 𝗱𝗲𝗺𝗼 𝗳𝗼𝗿 𝗮 𝗹𝗶𝘃𝗲 𝗲𝘅𝗮𝗺𝗽𝗹𝗲.\n\n#ai #aiagent #automation #gmail #openai #workflow #emailautomation #aiagents #techstack #efficiency #coldemail #leadgen #B2BMarketing #LeadGeneration #MarketingAutomation #EmailMarketing #MarketingOps #DigitalMarketing #GrowthHacking #MarketingStrategy #MachineLearning #n8n #GenerativeAI #google #sales #cms #dxp #composable #chatgpt #marketing #blog #content #viralpost #blogcontent",
    "url": "https://www.linkedin.com/posts/jimdejager_ai-aiagent-automation-activity-7358063647660621824-gqY5?utm_source=social_share_send&utm_medium=member_desktop_web&rcm=ACoAAF3IbmsBapYWOUSHvOaplfrho6JFb1hIMXM",
    "post_type": "regular",
    "author": {
      "first_name": "Jim",
      "last_name": "de Jager",
      "headline": "Building AI Bots & Automations 🏆➤ Consultative Sales | Solution Sales | Presales | Customer Success & Change Management",
      "username": "jimdejager",
      "profile_url": "https://www.linkedin.com/in/jimdejager?miniProfileUrn=urn%3Ali%3Afsd_profile%3AACoAAAwYQP0BOafQGTJtNJDighWarr4fn-EvQkA",
      "profile_picture": "https://media.licdn.com/dms/image/v2/D4E35AQEkQbsQSGOprw/profile-framedphoto-shrink_800_800/B4EZflg2q5GwAk-/0/1751902284898?e=1756033200&v=beta&t=69CPt5S_6Vf7C6pR3ybUSy14_76FH8La3YIS9kWS1vo"
    },
    "stats": {
      "total_reactions": 8,
      "like": 7,
      "support": 0,
      "love": 0,
      "insight": 1,
      "celebrate": 0,
      "funny": 0,
      "comments": 1,
      "reposts": 0
    },
    "article": {
      "url": "https://www.linkedin.com/pulse/how-i-built-ai-agent-creates-posts-viral-linkedin-content-de-jager-ovfke?trackingId=bQw2NOj7eZHhPOD0xs59cg%3D%3D",
      "title": "How I built an AI Agent that creates and posts viral LinkedIn content",
      "subtitle": "Jim de Jager"
    },
    "profile_input": "https://www.linkedin.com/in/jimdejager/"
  },
  {
    "urn": "7357663155218440192",
    "full_urn": "urn:li:activity:7357663155218440192",
    "posted_at": {
      "date": "2025-08-03 08:46:34",
      "relative": "2 weeks ago • Visible to anyone on or off LinkedIn",
      "timestamp": 1754203594975
    },
    "text": "𝗔𝗹𝗯𝗲𝗿𝘁 𝗛𝗲𝗶𝗷𝗻 𝘃𝗲𝗿𝘀𝗽𝗶𝗹𝘁 𝟯𝟴% 𝗺𝗶𝗻𝗱𝗲𝗿 𝗺𝗲𝘁 𝗔𝗜 (en hoe jij hetzelfde kan bereiken)\n\nJe staat voor een halfvolle koelkast maar zonder kookinspiratie?\n\nJe bent niet de enige. Dit alledaagse dilemma leidt jaarlijks tot €2,5 miljard aan voedselverspilling in Nederland.\n\nAlbert Heijn heeft dit probleem stilletjes aangepakt met behulp van hun AI innovatie:  𝗦𝗰𝗮𝗻 & 𝗞𝗼𝗼𝗸.\n\n𝗪𝗮𝘁 𝗵𝗲𝗯𝗯𝗲𝗻 𝘇𝗲 𝗴𝗲𝗱𝗮𝗮𝗻?\n- De oprichting van 𝗚𝗲𝗻 𝗔𝗜 𝗟𝗮𝗯𝘀 als interne AI-start-up\n- De ontwikkeling van technologie die foto's van ingrediënten analyseert\n- De integratie van 𝗔𝗜 𝗮𝘀𝘀𝗶𝘀𝘁𝗲𝗻𝘁 𝗦𝘁𝗲𝗶𝗷𝗻 in de app om slimmer boodschappen te doen\n\n𝗪𝗮𝘁 𝗹𝗲𝘃𝗲𝗿𝘁 𝗵𝗲𝘁 𝗼𝗽?\n- 𝟯𝟴% 𝗺𝗶𝗻𝗱𝗲𝗿 𝘃𝗼𝗲𝗱𝘀𝗲𝗹𝘃𝗲𝗿𝘀𝗽𝗶𝗹𝗹𝗶𝗻𝗴 bij huishoudens die de functie gebruiken\n- 𝟮𝟮% 𝗵𝗼𝗴𝗲𝗿𝗲 𝗸𝗹𝗮𝗻𝘁𝘁𝗲𝘃𝗿𝗲𝗱𝗲𝗻𝗵𝗲𝗶𝗱\n- Een duidelijke stap richting hun doel: 𝟱𝟬% 𝗺𝗶𝗻𝗱𝗲𝗿 𝘃𝗲𝗿𝘀𝗽𝗶𝗹𝗹𝗶𝗻𝗴 𝗶𝗻 𝟮𝟬𝟯𝟬\n\n𝗪𝗮𝘁 𝗸𝗮𝗻 𝗷𝗶𝗷 𝗵𝗶𝗲𝗿𝘃𝗮𝗻 𝗹𝗲𝗿𝗲𝗻?\nVeel automatiseringsprojecten mislukken vanwege de omvang of vaagheid. Albert Heijn koos een scherp omlijnd meetbaar probleem: voedselverspilling als gevolg van gebrek aan maaltijdplanning.\n\nZe pasten drie essentiële AI principes toe die elk bedrijf kan benutten:\n\n𝗣𝗿𝗼𝗯𝗹𝗲𝗲𝗺𝗴𝗲𝗿𝗶𝗰𝗵𝘁𝗲 𝗶𝗻𝘇𝗲𝘁 𝘃𝗮𝗻 𝗔𝗜\nRicht je op een concreet en herkenbaar pijnpunt en niet op 'algemene efficiëntie'.\n\n𝗚𝗲𝗯𝗿𝘂𝗶𝗸𝘀𝗴𝗲𝗺𝗮𝗸 𝘀𝘁𝗮𝗮𝘁 𝘃𝗼𝗼𝗿𝗼𝗽\nLaat gebruikers simpelweg een foto maken zonder ingewikkelde interfaces.\n\n𝗦𝗹𝘂𝗶𝘁 𝗮𝗮𝗻 𝗼𝗽 𝗯𝗲𝘀𝘁𝗮𝗮𝗻𝗱𝗲 𝗴𝗲𝘄𝗼𝗼𝗻𝘁𝗲𝗻\nIntegreer AI in kanalen en tools die klanten al gebruiken.\n\n𝗪𝗲𝗹𝗸 𝘀𝗽𝗲𝗰𝗶𝗳𝗶𝗲𝗸 𝗽𝗿𝗼𝗯𝗹𝗲𝗲𝗺 𝗸𝗮𝗻 𝗷𝗼𝘂𝘄 𝗼𝗿𝗴𝗮𝗻𝗶𝘀𝗮𝘁𝗶𝗲 𝗼𝗽𝗹𝗼𝘀𝘀𝗲𝗻 𝗺𝗲𝘁 𝗱𝗼𝗲𝗹𝗴𝗲𝗿𝗶𝗰𝗵𝘁𝗲 𝗔𝗜?\n\nDeel het hieronder!\n\n#ai #aiagent #automation #gmail #openai #workflow #emailautomation #aiagents #techstack #efficiency #coldemail #leadgen #B2BMarketing #LeadGeneration #MarketingAutomation #EmailMarketing #MarketingOps #DigitalMarketing #GrowthHacking #MarketingStrategy #MachineLearning #GenerativeAI #google #sales #cms #dxp #composable #chatgpt #AlbertHeijn #supermarkt #voedselverspilling #FMCG #retail",
    "url": "https://www.linkedin.com/posts/jimdejager_ai-aiagent-automation-activity-7357663155218440192-Alw8?utm_source=social_share_send&utm_medium=member_desktop_web&rcm=ACoAAF3IbmsBapYWOUSHvOaplfrho6JFb1hIMXM",
    "post_type": "regular",
    "author": {
      "first_name": "Jim",
      "last_name": "de Jager",
      "headline": "Building AI Bots & Automations 🏆➤ Consultative Sales | Solution Sales | Presales | Customer Success & Change Management",
      "username": "jimdejager",
      "profile_url": "https://www.linkedin.com/in/jimdejager?miniProfileUrn=urn%3Ali%3Afsd_profile%3AACoAAAwYQP0BOafQGTJtNJDighWarr4fn-EvQkA",
      "profile_picture": "https://media.licdn.com/dms/image/v2/D4E35AQEkQbsQSGOprw/profile-framedphoto-shrink_800_800/B4EZflg2q5GwAk-/0/1751902284898?e=1756033200&v=beta&t=69CPt5S_6Vf7C6pR3ybUSy14_76FH8La3YIS9kWS1vo"
    },
    "stats": {
      "total_reactions": 3,
      "like": 3,
      "support": 0,
      "love": 0,
      "insight": 0,
      "celebrate": 0,
      "funny": 0,
      "comments": 1,
      "reposts": 0
    },
    "media": {
      "type": "image",
      "url": "https://media.licdn.com/dms/image/v2/D4E22AQFnFpciaUO1xw/feedshare-shrink_1280/B4EZhurpbfHEAk-/0/1754203593500?e=1758153600&v=beta&t=MzHz0XEbnK2tnsZYaK6P9eVeenF1x5QZl18BkOsziRM",
      "images": [
        {
          "url": "https://media.licdn.com/dms/image/v2/D4E22AQFnFpciaUO1xw/feedshare-shrink_1280/B4EZhurpbfHEAk-/0/1754203593500?e=1758153600&v=beta&t=MzHz0XEbnK2tnsZYaK6P9eVeenF1x5QZl18BkOsziRM",
          "width": 1024,
          "height": 1536
        }
      ]
    },
    "profile_input": "https://www.linkedin.com/in/jimdejager/"
  },
  {
    "timestamp": "2025-08-17T10:25:04.122Z",
    "summary": {
      "totalProfiles": 1,
      "successfulProfiles": 1,
      "failedProfiles": 0,
      "totalPostsExtracted": 3,
      "failedUsernames": []
    },
    "results": [
      {
        "username": "https://www.linkedin.com/in/jimdejager/",
        "success": true,
        "postsCount": 3,
        "responseStatus": 200
      }
    ]
  }
]

here the logg from the succesfully run in Apify.
Why are the values still incorrect in the airtable node create a record?

[Wouter_Nigrini]

My wild guess is that youre trying to add values from a node which is NOT immediately before the “Create a record” node. When you use $json, it only looks at values from the direct previous node. Try changing “$json” to the explicit node name from where you got that result from above. I would assume it’s the node called “Apify - Scrape LinkedIn Posts” so use `$(’Apify - Scrape LinkedIn Posts’).first().json`

[jimvlieger]

Hey Wouter, I believe I’ve found the solution. I’ve added an extra URL input in the Airtable field to make scraping data from the LinkedIn account easier

Airtable

Airtable node

In the ‘Create a Record’ node, I’ve added the input values from the Apify node. Because of the URL link, it will now receive the correct data

The result is an overview of the scraped LinkedIn post, complete with a rewritten example for immediate use on LinkedIn. If your competitors have viral content, you can now easily leverage their success

The only annoying thing is that the Apify app keeps scraping 88+ results, even though I’ve set the maximum to 3. I’ve sent a ticket to the Apify help desk since the problem is within their application. I will hold off on testing with multiple LinkedIn creators until I can confirm that the scraping settings are under control, as the current volume is causing me to use a significant number of tokens

[Wouter_Nigrini]

That looks better

[jimvlieger]

Thanks for the support!

[jimvlieger]

@Wouter_Nigrini I’ve set up the Apify console correctly. The last 3 items of all LinkedIn profiles are being scraped. However, I don’t understand why a factor of 3 is being used in n8n. As you can see in the screenshot, exactly 3 messages per Linkedin profile are being scraped, but then multiplied by 3. Do you know which node I can adjust this in?